WO2019047327A1 - 一种智能球、***及方法 - Google Patents

一种智能球、***及方法 Download PDF

Info

Publication number
WO2019047327A1
WO2019047327A1 PCT/CN2017/105718 CN2017105718W WO2019047327A1 WO 2019047327 A1 WO2019047327 A1 WO 2019047327A1 CN 2017105718 W CN2017105718 W CN 2017105718W WO 2019047327 A1 WO2019047327 A1 WO 2019047327A1
Authority
WO
WIPO (PCT)
Prior art keywords
axis
angular velocity
acceleration
cal
triaxial
Prior art date
Application number
PCT/CN2017/105718
Other languages
English (en)
French (fr)
Inventor
吴建成
韩步勇
罗向望
张也雷
吕柏翰
Original Assignee
简极科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 简极科技有限公司 filed Critical 简极科技有限公司
Publication of WO2019047327A1 publication Critical patent/WO2019047327A1/zh

Links

Images

Classifications

    • 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
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • A—HUMAN NECESSITIES
    • A63—SPORTS; GAMES; AMUSEMENTS
    • A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
    • A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
    • A63B71/06—Indicating or scoring devices for games or players, or for other sports activities

Definitions

  • the invention relates to the field of intelligent sports equipment, in particular to a smart ball, system and method.
  • smart devices are also used in sports, such as the patent CN102779319A, and the intelligent monitoring system is used for outdoor sports, but the intelligent system only implements information sharing and navigation positioning modules of various clients in the regional network, and is mainly used in modules. Single or multiple outdoor athletes.
  • the main object of the present invention is to overcome the above drawbacks in the prior art and to propose a smart ball and system.
  • the system and the method provide a new relationship between the acceleration variable, the angular velocity variable and the centripetal force, and obtain more accurate triaxial angular velocity and triaxial angular velocity, so as to provide a basis for improving the athlete's special ability.
  • a data processing method for a smart ball which is characterized in that a three-axis acceleration variable and a three-axis angular velocity variable of a sphere are detected by a six-axis inertial sensor; first, a sensor signal is used to correct a neural network to realize a three-axis acceleration variable and a three-axis angular velocity variable. The signal correction is processed, and the centripetal force correction neural network is used to calculate the acceleration generated by the centripetal force; finally, the actual triaxial acceleration and the actual triaxial angular velocity are calculated according to the acceleration generated by the centripetal force, the corrected triaxial acceleration, and the triaxial angular velocity.
  • the sensor signal correction neural network comprises three layers, the input layer comprises the triaxial acceleration and the triaxial angular velocity, and the output layer is the corrected triaxial acceleration and the triaxial angular velocity, and the neurons in the hidden layer are activated.
  • the function is an S function, and the number of neurons in the hidden layer is 100.
  • the sensor signal correction neural network is trained by a back propagation algorithm, and the training set collects data at different azimuths and elevation angles through a three-axis correction dial, and the azimuth angle and the elevation angle range from 0 degrees to 360 degrees, the six axes.
  • the inertial sensor is at a distance from the center of the three-axis correction dial, and its x-axis is positively directed toward the center of the three-axis correction dial.
  • the centripetal force correction neural network comprises three layers, the input layer is the corrected triaxial angular velocity, the output layer is the acceleration generated by the centripetal force, the neuron activation function in the hidden layer is the S function, and the nerve in the hidden layer The number of elements is 100.
  • the centripetal force correction neural network is trained by a back propagation algorithm, and the training set collects data at different azimuths, elevation angles and rotation speeds through a three-axis correction dial, and the azimuth angle and the elevation angle range are from 0 degrees to 360 degrees, and the rotation speed of the turntable The range is from 0 rpm to 300 rpm; the six-axis inertial sensor is at a distance from the center of the three-axis correction dial, and its x-axis is positively directed toward the center of the three-axis correction dial.
  • the actual triaxial acceleration and triaxial angular velocity are calculated according to the following formula:
  • a [a x , a y , a z ] is the calculated actual triaxial acceleration
  • w [w x , w y , w z ] is the calculated actual triaxial angular velocity
  • a mea [a mea , x , a mea, y , a mea, z ] is the detected triaxial acceleration
  • a cal [a cal,x , a cal,y , a cal,z ] is the corrected triaxial acceleration
  • f cal (.) corrects the neural network model for the sensor signal
  • c [c x , c y , c z ] is the acceleration generated by the centripetal force caused by the rotation
  • w mea [w mea, x , w mea, y , w Mea,z ] is the detected three-axis angular velocity
  • ⁇ cal [ ⁇ cal,x , ⁇ cal,y
  • a smart ball comprising a sphere, further comprising a micro control unit mounted on the sphere, an inertial sensing device and a wireless communication device mounted on the spherical sphere and connected to the micro control unit, including a six-axis inertial sensor and a sensor esProc; the six-axis inertial sensor is coupled to the sensor esProc to detect acceleration variables and angular velocity variables; the sensor calculator uses data of any of the above-mentioned smart balls The processing method calculates the actual triaxial acceleration and the triaxial angular velocity and sends them to the micro control unit through the wireless communication device.
  • a wireless charging receiving device and a battery device are further included, the battery device being connected to the respective devices to provide power, and the wireless charging receiving device is connected to the battery device for wirelessly charging the battery device.
  • a smart ball system comprising: any one of the above-mentioned smart balls, mobile devices and servers, wherein the smart balls communicate with mobile devices and servers via wireless communication devices.
  • a smart ball system comprising: any one of the above-mentioned smart balls, a plurality of UWB positioning base stations, a gateway, and a server, wherein the smart ball is provided with a UWB wireless transceiver device to realize positioning of the base station with the UWB Data communication between the UWB positioning base station transmits data to the gateway through wireless or wired means, and the gateway forwards the data to the server through wireless or wired means.
  • the present invention has the following advantageous effects as compared with the prior art:
  • the smart ball and method of the present invention uses a sensor signal correction neural network to realize pre-processing signal correction of a triaxial acceleration variable and a triaxial angular velocity variable, and combines a centripetal force correction neural network to calculate an acceleration generated by centripetal force; finally, according to an acceleration generated by centripetal force,
  • the corrected triaxial acceleration and triaxial angular velocity calculate the actual triaxial acceleration and triaxial angular velocity, providing a new relationship between acceleration variables, angular velocity variables and centripetal force, thus providing a more accurate data basis for improving the athlete's special ability.
  • the sensor signal correction neural network and the centripetal force correction neural network use a back propagation training algorithm to calculate and train the correction value, and filter a large number of training numerical classes through a mathematical model of the neural network to ensure The calculation results are accurate and reliable.
  • the smart ball system of the present invention combined with the mobile device and the server, further processes the motion-related data such as the actual triaxial angular velocity and the triaxial angular velocity obtained by the pre-processing into the training data for presentation by the mobile device, and can also be synchronized to The server performs storage;
  • the smart ball system of the present invention is realized by combining the UWB positioning base station, the gateway and the server, and realizing the calculation of the sports ball positioning and the motion track and the like by the UWB positioning base station, and providing a data basis for improving the athlete's special ability.
  • FIG. 1 is a schematic diagram of a sensor signal correction neural network of the present invention
  • FIG. 2 is a schematic diagram of a centripetal force correction neural network of the present invention
  • Figure 3 is a block diagram of the composition of the smart ball of the present invention.
  • Figure 4 is a block diagram of the composition of the smart ball system of the present invention.
  • Fig. 5 is a block diagram showing another composition of the smart ball system of the present invention.
  • a smart ball data processing method detects a three-axis acceleration variable, a three-axis angular velocity variable, a rotational speed, and the like of a sphere through a six-axis inertial sensor 12, and the six-axis inertial sensor 12 includes a three-axis angular velocity sensor and a three-axis angular velocity sensor.
  • the sensor signal is used to correct the neural network to realize the pre-processing signal correction of the triaxial acceleration variable and the triaxial angular velocity variable, and then the centripetal force correction neural network is used to calculate the acceleration generated by the centripetal force.
  • the corrected triaxial acceleration and the triaxial angular velocity the actual triaxial angular velocity and triaxial angular velocity are calculated according to the following formula:
  • a [a x , a y , a z ] is the calculated actual triaxial acceleration
  • w [w x , w y , w z ] is the calculated actual triaxial angular velocity
  • a mea [a mea , x , a mea, y , a mea, z ] is the detected triaxial acceleration
  • a cal [a cal,x , a cal,y , a cal,z ] is the corrected triaxial acceleration
  • f cal (.) corrects the neural network model for the sensor signal
  • c [c x , c y , c z ] is the acceleration generated by the centripetal force caused by the rotation.
  • w mea [w mea,x ,w mea,y ,w mea,z ] is the detected triaxial angular velocity
  • ⁇ cal [ ⁇ cal,x , ⁇ cal,y , ⁇ cal,z ]
  • the post-three-axis angular velocity is also the calculated actual triaxial angular velocity
  • f cent (.) is the centripetal force corrected neural network model.
  • the sensor signal correction neural network of the present invention is trained by a back propagation algorithm, first defined For the weight value of the i-th neuron of the mth layer and the jth neuron of the nth layer, the best weight value is iteratively calculated by using a back propagation algorithm:
  • k is the number of iterations and the highest iteration number is set to 50
  • ⁇ is the momentum constant and is set to 0.5
  • ⁇ is the learning rate and is set to 0.01
  • E is the loss function and is set to the square loss function.
  • the training set collects relevant data at different azimuths and elevation angles through a three-axis correction dial.
  • the azimuth angle and elevation angle range from 0 degrees to 360 degrees with an interval of 1 degree.
  • the six-axis inertial sensor 12 is placed at a distance from the center of the three-axis correction dial, which is fixed at 110 mm, and the x-axis of the three-axis angular velocity sensor and the three-axis angular velocity sensor are directed to the center of the three-axis correction dial.
  • the centripetal force correction neural network is used to provide a relationship between triaxial rotation and triaxial force.
  • the neuron h 1 , h 2 ,...h 100 activation function is the S function: The number of neurons in the hidden layer is 100.
  • centripetal force correction neural network of the present invention is trained by a back propagation algorithm, and the sensor signal is corrected with the aforementioned neural network, and the back weight propagation algorithm is used to iteratively calculate the optimal weight value.
  • k is the number of iterations and the highest iteration number is set to 80
  • ⁇ is the momentum constant and is set to 0.5
  • ⁇ is the learning rate and is set to 0.05
  • E is the loss function and is set to the square loss function.
  • the training set collects relevant data at different azimuths, elevation angles and rotational speeds through a three-axis correction dial.
  • the azimuth angle and elevation angle range from 0 degrees to 360 degrees, and the angle interval is 1 degree.
  • the turntable speed ranged from 0 rpm to 300 rpm with a rotational speed interval of 5 rpm.
  • the six-axis inertial sensor 12 is spaced from the center of the three-axis correction dial by a distance of 110 mm, and the x-axis of the three-axis angular velocity sensor and the three-axis angular velocity sensor are directed toward the center of the three-axis correction dial.
  • the present invention further provides a smart ball, including a ball 10, a micro control unit 11 mounted on the ball 10, an inertial sensing device, a clock device 14, a wireless charging receiving device 15, a storage device 16, and a battery device 17. And a wireless communication device.
  • the inertial sensing device is mounted on the ball of the sphere and includes a six-axis inertial sensor 12 and a sensor esProc 13.
  • the six-axis inertial sensor 12 is connected to the sensor esProc 13 to detect an acceleration variable, an angular velocity variable, a rotating shaft, a rotational speed, and the like, and employs a three-axis acceleration sensor and a three-axis angular velocity sensor that can measure the sphere 10
  • the center of gravity tilts, moves up and down, left and right, and changes in the movement in space; the three-axis angular velocity sensor uses the physical force caused by the Coriolis force principle to measure the angular velocity variable of each axis.
  • the sensor esProc 13 is connected to the micro control unit 11 for three-axis angular velocity and three-axis acceleration
  • the variables are preprocessed by the above data processing method to obtain motion related data, including actual triaxial acceleration, triaxial angular velocity, etc., and then sent to the micro control unit 11.
  • the sensor esProc 13 can also reduce power consumption, share the data processing task of the micro control unit 11, embed data storage, and shorten the wake-up time of the micro control unit 11.
  • the micro control unit 11 is connected to the wireless communication device for combining motion related data and a clock signal and transmitting the same through a wireless communication device, and the micro control unit can be implemented by using a single chip microcomputer.
  • the wireless communication device of the present invention includes a Bluetooth unit 19 and a WIFI unit 18, which can be used to implement data communication with the mobile device 30, including transmitting motion-related data containing timestamps and receiving control commands including setting the ball Name, set data collection period, set connection, etc.
  • the WIFI unit 18 implements data communication with a mobile device or server, including transmitting motion-related data containing time stamps to a mobile device or server, and receiving control commands from the server or mobile device 30.
  • the sphere 10 of the present invention is basketball or soccer or volleyball or rugby or handball, etc., and its working principle is as follows:
  • the six-axis inertial sensor 12 in the sphere 10 collects motion-related data of the sphere 10 in real time, including triaxial acceleration, triaxial angular velocity, rotational axis and rotational speed, etc., and is intelligently performed by the sensor esProc 13
  • the ball data processing obtains relevant motion information, which is then sent to the micro control unit 11, which, in conjunction with the clock information from the clock device 14, adds a timestamp to the motion related data and transmits it via the wireless communication device.
  • the present invention also provides a smart ball system comprising the above described smart ball, mobile device 30 and server 20, the sphere 10 of which is in data communication with the mobile device 30 and server 20 via a wireless communication device.
  • the mobile device 30 can be a smart terminal having a wireless communication function, such as a mobile phone, a tablet, a watch, a notebook, etc., which receives motion-related data with a time stamp, and performs processing to calculate a trajectory of the movement of the sphere 10, and the trajectory is redirected by an external force.
  • the force is further converted into training data for presentation, and can also be synchronized to the server 20 for storage.
  • the present invention further proposes another smart ball system, including the above-mentioned smart ball, a plurality of UWB positioning base stations 50, a gateway 60, and a server 20.
  • the plurality of UWB positioning base stations 50 are disposed at specific corners of the stadium, and the number thereof may be three or four, and the sphere 10 of the smart ball is further provided with a UWB radio transceiver to implement data communication with the UWB positioning base station 50.
  • the UWB positioning base station 50 transmits the time-stamped motion-related data and the data arrival time to the gateway 60 by wireless or wired means, and the gateway 60 forwards the data to the server 20 by wireless or wired, and the server 20 passes the data.
  • the sphere 10 is positioned to reach the time required by each base station or the time difference of arrival, etc., and motion data such as a motion trajectory is obtained.
  • the server 20 may include a cloud server or a local server, which is a server that analyzes the motion trajectory of the player and the smart ball, and does not store all the analysis results.
  • the function of the cloud server is to store the results of data analysis and synchronize the user's data, but only a small amount of data analysis.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Navigation (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种智能球、***及方法被公开。该智能球***通过六轴惯性传感器（12）检测球体（10）的三轴加速度变量和三轴角速度变量，先采用传感器信号校正神经网络实现三轴加速度变量与三轴角速度变量的预处理信号校正，再采用向心力校正神经网络计算因向心力产生的加速度，最后根据向心力产生的加速度、校正后的三轴加速度和三轴角速度计算实际的三轴加速度和实际的三轴角速度。该智能球提供了新的加速度变量、角速度变量与向心力之间的关系，从而为提高运动员专项能力提供更为准确的数据依据。

Description

一种智能球、***及方法 技术领域
本发明涉及智能化运动设备领域，特别是一种智能球、***及方法。
背景技术
近年来，智能设备产业发展迅猛。智能设备受到如此追捧的一个主要原因在于，面向智能设备的软件应用开发与发布生态***的形成。每天都有数以百计的新应用流入市场，不断促进着该生态***的进化。但是，在这数量庞大的软件应用中，以纯软件或网页应用为主流，对智能设备内部的各类传感器(重力传感器、陀螺仪、磁力计、全球定位***等)的使用仅限于地图和游戏等。
在现有技术中也有将智能设备运用于运动中，如专利CN102779319A，将智能监控***用于户外运动，但是该智能***只是在区域网络实现各个客户端的信息共享及导航定位等模块，主要运用于单个或多个户外运动者。
对于足球、篮球或排球等高强度、大负荷的比赛项目，要求运动员在跑动中，通过准确发挥个人和集体的战术水平，达到战胜对手的目的。因此，如何将高强度训练负荷与运动员的专项技术高度融合，是提高运动员专项能力的关键。目前，主要是利用足球的整体受力表现进一步的计算分析进阶数据。然而现有技术大多是通过足球内的六轴惯性传感器检测足球整体的加速度变量与角速度变量并且建立一个误差模型或方程从中计算足球质心受力,然而其效能则受限于对应关系的复杂度,多变量与非线性关系。
发明内容
本发明的主要目的在于克服现有技术中的上述缺陷，提出一种智能球、系 统及方法，提供了新的加速度变量、角速度变量与向心力之间的关系，获得更为准确的三轴角速度和三轴角速度，从而更好地为提高运动员专项能力提供依据。
本发明采用如下技术方案：
一种智能球的数据处理方法，其特征在于：通过六轴惯性传感器检测球体的三轴加速度变量和三轴角速度变量；先采用传感器信号校正神经网络实现三轴加速度变量与三轴角速度变量的预处理信号校正，再采用向心力校正神经网络计算向心力产生的加速度；最后根据向心力产生的加速度、校正后的三轴加速度和三轴角速度计算实际的三轴加速度和实际的三轴角速度。
优选的，所述传感器信号校正神经网络包括三层，其输入层包括所述三轴加速度和三轴角速度,输出层则为校正后的三轴加速度和三轴角速度,隐藏层中的神经元激活函数为S函数,隐藏层中的神经元数目为100。
优选的，所述传感器信号校正神经网络采用反向传播算法训练，训练集通过三轴校正转盘在不同方位角与仰角收集数据，方位角度与仰角角度范围为0度到360度，所述六轴惯性传感器距离该三轴校正转盘中心一定距离，其x轴正向指向该三轴校正转盘中心。
优选的，所述向心力校正神经网路包括三层,其输入层为校正后的三轴角速度,输出层为向心力产生的加速度,隐藏层中的神经元激活函数为S函数,隐藏层中的神经元数目为100。
优选的，所述向心力校正神经网络采用反向传播算法训练，其训练集通过三轴校正转盘在不同方位角、仰角和转速收集数据，方位角度与仰角角度范围为0度到360度，转盘转速范围为0rpm到300rpm；所述六轴惯性传感器距离该三轴校正转盘中心一定距离，其x轴正向指向该三轴校正转盘中心。
优选的，根据下式计算实际的三轴加速度和三轴角速度：
a＝acal-c,
Figure PCTCN2017105718-appb-000001
c＝fcent(ωcal)
其中：a＝[ax,ay,az]为计算后的实际三轴加速度，w＝[wx,wy,wz]为计算后的实际三轴角速度，amea＝[amea,x,amea,y,amea,z]为检测到的所述三轴加速度，acal＝[acal,x,acal,y,acal,z]为校正后的三轴加速度，fcal(.)为传感器信号校正神经网络模型，c＝[cx,cy,cz]为旋转造成的向心力所产生的加速度；wmea＝[wmea,x,wmea,y,wmea,z]为检测到的所述三轴角速度，ωcal＝[ωcal,x,ωcal,y,ωcal,z]为校正后的三轴角速度，亦为计算后的实际三轴角速度，fcent(.)为向心力校正神经网络模型。
一种智能球，包括球体，还包括安装于球体上的微控制单元、惯性传感装置和无线通信装置，该惯性传感装置安装于所述球体球皮上且与微控制单元相连，其包括六轴惯性传感器和传感集算器；其特征在于：该六轴惯性传感器与传感集算器相连以检测加速度变量和角速度变量；该传感计算器采用上述的任意一种智能球的数据处理方法来计算实际的三轴加速度和三轴角速度，并通过无线通信装置送至微控制单元。
优选的，还包括无线充电接收装置和电池装置，该电池装置与上述各个装置相连以提供电源，该无线充电接收装置与电池装置相连用于对电池装置进行无线充电。
一种智能球***，其特征在于：包括上述的任一一种智能球、移动设备和服务器，该智能球通过无线通信装置与移动设备和服务器实现数据通信。
一种智能球***，其特征在于：包括上述的任一一种智能球、若干UWB定位基站、网关、服务器，该智能球设有UWB无线收发装置以实现与UWB定位基站 之间的数据通信；该UWB定位基站通过无线或有线方式将数据传输至网关，该网关通过无线或有线方式将数据转发至服务器。
由上述对本发明的描述可知，与现有技术相比，本发明具有如下有益效果：
1、本发明的智能球和方法，采用传感器信号校正神经网络实现三轴加速度变量与三轴角速度变量的预处理信号校正，结合向心力校正神经网络计算向心力产生的加速度；最后根据向心力产生的加速度、校正后的三轴加速度和三轴角速度计算实际的三轴加速度和三轴角速度，提供了新的加速度变量、角速度变量与向心力之间的关系，从而为提高运动员专项能力提供更为准确的数据依据。
2、本发明的智能球和方法，传感器信号校正神经网络和向心力校正神经网络采用反向传播训练算法计算并训练得出校正值，通过类神经网络的数学模型筛选大量的训练数值类，从而确保计算结果的准确和可靠。
3、本发明的智能球***，结合移动设备和服务器，通过移动设备将预处理得到的实际三轴角速度、三轴角速度等运动相关的数据进行进一步处理转化成训练数据进行呈现，还可同步至服务器进行存储；
4、本发明的智能球***，其结合UWB定位基站、网关和服务器实现，通过UWB定位基站实现智能球的定位及运动轨迹等运动数据的计算，为提高运动员专项能力提供数据依据。
附图说明
图1为本发明的传感器信号校正神经网络示意图；
图2为本发明的向心力校正神经网络示意图；
图3为本发明智能球的组成框图；
图4为本发明智能球***的组成框图；
图5为本发明智能球***的另一组成框图。
其中：10、球体，11、微控制单元，12、六轴惯性传感器，13、传感集算器吗，14、时钟装置，15、无线充电接收装置、16、存储装置，17、电池装置，18、WIFI单元，19、蓝牙单元，20、服务器，30、移动设备,50、UWB定位基站，60、网关。
具体实施方式
以下通过具体实施方式对本发明作进一步的描述。
一种智能球的数据处理方法，通过六轴惯性传感器12检测球体的三轴加速度变量、三轴角速度变量、旋转速度等，该六轴惯性传感器12包括三轴角速度传感器和三轴角速度传感器。先采用传感器信号校正神经网络实现三轴加速度变量与三轴角速度变量的预处理信号校正，再采用向心力校正神经网络计算向心力产生的加速度。最后根据向心力产生的加速度、校正后的三轴加速度和三轴角速度，根据下式计算实际的三轴角速度和三轴角速度：
a＝acal-c,
Figure PCTCN2017105718-appb-000002
c＝fcent(ωcal)
其中：a＝[ax,ay,az]为计算后的实际三轴加速度，w＝[wx,wy,wz]为计算后的实际三轴角速度，amea＝[amea,x,amea,y,amea,z]为检测到的所述三轴加速度，acal＝[acal,x,acal,y,acal,z]为校正后的三轴加速度，fcal(.)为传感器信号校正神经网络模型，c＝[cx,cy,cz]为旋转造成的向心力所产生的加速度。
wmea＝[wmea,x,wmea,y,wmea,z]为检测到的所述三轴角速度，ωcal＝[ωcal,x,ωcal,y,ωcal,z]为校正后的三轴角速度，亦为计算后的实际三轴角速度，fcent(.)为向心力校正神经 网络模型。
参照图1，传感器信号校正神经网络用于校正三轴加速度传感器和三轴角速度传感器，该神经网络包括三层，其输入层i1,i2,...i6包括三轴加速度amea＝[amea,x,amea,y,amea,z]和三轴角速度wmea＝[wmea,x,wmea,y,wmea,z],输出层o1,o2,...o6则为修正后的三轴加速度acal＝[acal,x,acal,y,acal,z]和三轴角速度ωcal＝[ωcal,x,ωcal,y,ωcal,z],隐藏层h1,h2,...h100中的神经元激活函数lact(.)为S函数
Figure PCTCN2017105718-appb-000003
隐藏层中的神经元数目为100。该修正后的三轴角速度ωcal即为实际的三轴角速度w＝[wx,wy,wz]。
本发明的该传感器信号校正神经网络采用反向传播算法训练，首先定义
Figure PCTCN2017105718-appb-000004
为第m层第i个神经元与第n层第j个神经元连接的权重值，利用反向传播算法迭代计算出最佳权重值：
Figure PCTCN2017105718-appb-000005
Figure PCTCN2017105718-appb-000006
其中：k为迭代次数并设定最高迭代次数为50，μ为动量常数并设定为0.5，α为学***方损失函数。
训练集通过三轴校正转盘在不同方位角与仰角收集相关数据，方位角度与仰角角度范围为0度到360度，间隔为1度。六轴惯性传感器12放置于距离该三轴校正转盘中心一定距离的位置，该距离固定为110mm，其三轴角速度传感器和三轴角速度传感器的x轴正向指向该三轴校正转盘中心。
参照图2，向心力校正神经网路用于提供三轴旋转与三轴向心力的关系，其神经网络包括三层,其输入层i1,i2,i3为校正后的三轴角速度ωcal＝[ωcal,x,ωcal,y,ωcal,z],输出层o1,o2,o3为因向心力产生的加速度c＝[cx,cy,cz], 隐藏层中的神经元h1,h2,...h100激活函数为S函数:
Figure PCTCN2017105718-appb-000007
隐藏层中的神经元数目为100。
本发明的向心力校正神经网络采用反向传播算法训练，与前述的传感器信号校正神经网络，利用反向传播算法迭代计算出最佳权重值
Figure PCTCN2017105718-appb-000008
Figure PCTCN2017105718-appb-000009
Figure PCTCN2017105718-appb-000010
其中：k为迭代次数并设定最高迭代次数为80，μ为动量常数并设定为0.5，α为学***方损失函数。其训练集通过三轴校正转盘在不同方位角、仰角和转速收集相关数据，方位角度与仰角角度范围为0度到360度，角度间隔为1度。转盘转速范围为0rpm到300rpm，转速间隔为5rpm。六轴惯性传感器12距离该三轴校正转盘中心一定距离，该距离固定为110mm，其三轴角速度传感器和三轴角速度传感器的x轴正向指向该三轴校正转盘中心。
参照图3，本发明还提出一种智能球，包括球体10及安装于球体10上的微控制单元11、惯性传感装置、时钟装置14、无线充电接收装置15、存储装置16、电池装置17和无线通信装置。该惯性传感装置安装于球体的球皮上，且包括六轴惯性传感器12和传感集算器13。
该六轴惯性传感器12与传感集算器13相连以检测加速度变量、角速度变量、旋转轴和旋转速度等，其采用三轴加速度传感器和三轴角速度传感器，该三轴加速度传感器可测量球体10重心倾斜、上下左右晃动以及空间中的移动变化；该三轴角速度传感器利用科里奥利力原理造成的物理作用力，测量各轴的角速度变量。
该传感集算器13与微控制单元11相连，用于对三轴角速度和三轴加速度 等变量，采用上述的数据处理方法进行预处理得到运动相关数据，包括实际的三轴加速度、三轴角速度等，再送至微控制单元11。另外，传感集算器13还能够降低功耗，分担了微控制单元11的数据处理任务，内嵌有数据存储，缩短微控制单元11的唤醒时间。
该微控制单元11与无线通信装置相连用于将运动相关数据和时钟信号结合并通过无线通信装置发送，该微控制单元可采用单片机实现。本发明的无线通信装置包括蓝牙单元19和WIFI单元18，该蓝牙单元19可用于实现与移动设备30的数据通信，包括发送含有时间戳的运动相关数据及接收控制命令，该控制命令包括设置球的名称、设置数据采集周期、设置连接等。该WIFI单元18实现与移动设备或服务器的数据通信，包括发送含有时间戳的运动相关数据至移动设备或服务器，及接收来自服务器或移动设备30的控制命令。
本发明的球体10为篮球或足球或排球或橄榄球或手球等，其工作原理如下：
当运动员进行训练时，球体10内的六轴惯性传感器12实时采集球体10的运动相关数据，包括三轴加速度、三轴角速度、旋转轴和旋转速度等，并经传感集算器13进行智能球数据处理得到相关的运动信息，后送至微控制单元11，微控制单元11结合来自时钟装置14的时钟信息，为运动相关数据添加时间戳，经无线通信装置进行发送。
参照图4，本发明还提出一种智能球***，包括上述的智能球、移动设备30和服务器20，该智能球的球体10通过无线通信装置与移动设备30和服务器20实现数据通信。该移动设备30可为手机、平板、手表、笔记本等具有无线通信功能的智能终端，其接收具有时间戳的运动相关数据，并进行处理后计算得到球体10运动的轨迹，及轨迹受外力改变方向的力，进一步转化成训练数据进行呈现，还可同步至服务器20进行存储。
参照图5，本发明又提出另一种智能球***，包括上述的智能球、若干UWB定位基站50、网关60、服务器20。该若干UWB定位基站50设置于球场的特定角落，其数量可以是三个或四个等，该智能球的球体10还设有UWB无线收发装置以实现与UWB定位基站50之间的数据通信。该UWB定位基站50通过无线或有线方式将包含有时间戳的与运动相关的数据、及数据到达时间传输至网关60，该网关60通过无线或有线方式将数据转发至服务器20，服务器20通过数据的到达各个基站所需的时间或到达时间差等对球体10进行定位，并得出运动轨迹等运动数据。该服务器20可包括云服务器或本地服务器，本地服务器是对球员和智能球的运动轨迹进行数据分析的服务器，并不会存储所有的分析结果。而云服务器的功能是存储数据分析的结果并且同步用户的数据，但只做少量的数据分析。
上述仅为本发明的具体实施方式，但本发明的设计构思并不局限于此，凡利用此构思对本发明进行非实质性的改动，均应属于侵犯本发明保护范围的行为。

Claims (10)

  1. 一种智能球的数据处理方法，其特征在于：通过六轴惯性传感器检测球体的三轴加速度变量和三轴角速度变量；先采用传感器信号校正神经网络实现三轴加速度变量与三轴角速度变量的预处理信号校正，再采用向心力校正神经网络计算向心力产生的加速度；最后根据向心力产生的加速度、校正后的三轴加速度和三轴角速度计算实际的三轴加速度和实际的三轴角速度。
  2. 如权利要求1所述的一种智能球的数据处理方法，其特征在于：所述传感器信号校正神经网络包括三层，其输入层包括所述三轴加速度和三轴角速度,输出层则为校正后的三轴加速度和三轴角速度,隐藏层中的神经元激活函数为S函数,隐藏层中的神经元数目为100。
  3. 如权利要求2所述的一种智能球的数据处理方法，其特征在于：所述传感器信号校正神经网络采用反向传播算法训练，训练集通过三轴校正转盘在不同方位角与仰角收集数据，方位角度与仰角角度范围为0度到360度，所述六轴惯性传感器距离该三轴校正转盘中心一定距离，其x轴正向指向该三轴校正转盘中心。
  4. 如权利要求1所述的一种智能球的数据处理方法，其特征在于：所述向心力校正神经网路包括三层,其输入层为校正后的三轴角速度,输出层为向心力产生的加速度,隐藏层中的神经元激活函数为S函数,隐藏层中的神经元数目为100。
  5. 如权利要求4所述的一种智能球的数据处理方法，其特征在于：所述向心力校正神经网络采用反向传播算法训练，其训练集通 过三轴校正转盘在不同方位角、仰角和转速收集数据，方位角度与仰角角度范围为0度到360度，转盘转速范围为0rpm到300rpm；所述六轴惯性传感器距离该三轴校正转盘中心一定距离，其x轴正向指向该三轴校正转盘中心。
  6. 如权利要求1所说的一种智能球的数据处理方法，其特征在于：根据下式计算实际的三轴加速度和三轴角速度：
    a＝acal-c,
    Figure PCTCN2017105718-appb-100001
    c＝fcent(ωcal)
    其中：a＝[ax,ay,az]为计算后的实际三轴加速度，w＝[wx,wy,wz]为计算后的实际三轴角速度，amea＝[amea,x,amea,y,amea,z]为检测到的所述三轴加速度，acal＝[acal,x,acal,y,acal,z]为校正后的三轴加速度，fcal(.)为传感器信号校正神经网络模型，c＝[cx,cy,cz]为旋转造成的向心力所产生的加速度；wmea＝[wmea,x,wmea,y,wmea,z]为检测到的所述三轴角速度，ωcal＝[ωcal,x,ωcal,y,ωcal,z]为校正后的三轴角速度，亦为计算后的实际三轴角速度，fcent(.)为向心力校正神经网络模型。
  7. 一种智能球，包括球体，还包括安装于球体上的微控制单元、惯性传感装置和无线通信装置，该惯性传感装置安装于所述球体球皮上且与微控制单元相连，其包括六轴惯性传感器和传感集算器；其特征在于：该六轴惯性传感器与传感集算器相连以检测加速度变量和角速度变量；该传感计算器采用权利要求1至6所述的任意一种智能球的数据处理方法来计算实际的三轴加速度和三轴角速度，并通过无线 通信装置送至微控制单元。
  8. 如权利要求7所述的一种智能球，其特征在于：还包括无线充电接收装置和电池装置，该电池装置与上述各个装置相连以提供电源，该无线充电接收装置与电池装置相连用于对电池装置进行无线充电。
  9. 一种智能球***，其特征在于：包括权利要求7所述的任一一种智能球、移动设备和服务器，该智能球通过无线通信装置与移动设备和服务器实现数据通信。
  10. 一种智能球***，其特征在于：包括权利要求7所述的任一一种智能球、若干UWB定位基站、网关、服务器，该智能球设有UWB无线收发装置以实现与UWB定位基站之间的数据通信；该UWB定位基站通过无线或有线方式将数据传输至网关，该网关通过无线或有线方式将数据转发至服务器。
PCT/CN2017/105718 2017-09-11 2017-10-11 一种智能球、***及方法 WO2019047327A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710812728.8 2017-09-11
CN201710812728.8A CN107679623B (zh) 2017-09-11 2017-09-11 一种智能球、***及方法

Publications (1)

Publication Number Publication Date
WO2019047327A1 true WO2019047327A1 (zh) 2019-03-14

Family

ID=61135266

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/105718 WO2019047327A1 (zh) 2017-09-11 2017-10-11 一种智能球、***及方法

Country Status (2)

Country Link
CN (1) CN107679623B (zh)
WO (1) WO2019047327A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108654041A (zh) * 2018-05-25 2018-10-16 姜哲 一种足球模拟训练服务器、终端、***及方法
CN109364456B (zh) * 2018-12-03 2023-09-29 北京汇洋时代科技有限公司 一种排球垫球测试装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008073210A (ja) * 2006-09-21 2008-04-03 Seiko Epson Corp ゴルフクラブ、そのスイング評価支援装置
CN101367013A (zh) * 2007-08-17 2009-02-18 阿迪达斯国际经营管理有限公司 具有运动球的运动电子训练***及其应用
CN102489007A (zh) * 2011-12-07 2012-06-13 深圳市江波龙电子有限公司 一种智能高尔夫球、终端***及数据采集分析方法
CN204680194U (zh) * 2015-06-25 2015-09-30 厦门市简极科技有限公司 一种智能球及其***
US20150283427A1 (en) * 2012-12-18 2015-10-08 Seiko Epson Corporation Motion analysis system and motion analysis method

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6560586B1 (en) * 1998-10-30 2003-05-06 Alcatel Multiresolution learning paradigm and signal prediction
CN205598601U (zh) * 2016-05-11 2016-09-28 东北石油大学 一种便携式运动数据捕捉装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008073210A (ja) * 2006-09-21 2008-04-03 Seiko Epson Corp ゴルフクラブ、そのスイング評価支援装置
CN101367013A (zh) * 2007-08-17 2009-02-18 阿迪达斯国际经营管理有限公司 具有运动球的运动电子训练***及其应用
CN102489007A (zh) * 2011-12-07 2012-06-13 深圳市江波龙电子有限公司 一种智能高尔夫球、终端***及数据采集分析方法
US20150283427A1 (en) * 2012-12-18 2015-10-08 Seiko Epson Corporation Motion analysis system and motion analysis method
CN204680194U (zh) * 2015-06-25 2015-09-30 厦门市简极科技有限公司 一种智能球及其***

Also Published As

Publication number Publication date
CN107679623B (zh) 2019-03-08
CN107679623A (zh) 2018-02-09

Similar Documents

Publication Publication Date Title
US20220241641A1 (en) Systems and Methods of Swimming Analysis
JP6454444B2 (ja) インテリジェントグランドシステム及びデータ取得方法
US9186568B2 (en) Basketball shot-tracking system
TW201702965A (zh) 智能球及其系統
CN105030246B (zh) 一种测量人体在运动中消耗能量的方法、装置及计步器
CN107250727A (zh) 确定移动设备在地理区域中的位置
CN104524749A (zh) 数字化智能篮球
CN104722058A (zh) 一种运动员竞技状态分析方法及***
CN103884337A (zh) 多姿态步距校正定位***与方法
CN204680194U (zh) 一种智能球及其***
CN110327595A (zh) 基于可穿戴传感器的动作捕捉识别与评估装置及方法
US11580878B2 (en) Real time sports motion training aid
CN106611153B (zh) 一种智能球训练动作识别***及方法
Straeten et al. Method and implementation of micro Inertial Measurement Unit (IMU) in sensing basketball dynamics
Gowda et al. Iot platform for sports analytics
WO2019047327A1 (zh) 一种智能球、***及方法
CN111208545B (zh) 一种运动轨迹记录方法与***
Shankar et al. Performance measurement and analysis of shooting form of basketball players using a wearable IoT system
US11273354B2 (en) Real time sports motion training aid
Abdallah et al. IoT device for Athlete's movements recognition using inertial measurement unit (IMU)
Chang et al. A low cost multi-sensors navigation solution for sport performance assessment
US20230149789A1 (en) System and method for golf super tag multifunction golf swing capture and analysis device
Bai et al. Using a wearable device to assist the training of the throwing motion of baseball players
Balbudhe et al. Smart Cricket Ball: Solutions to Design and Deployment Challenges
Wang et al. Guide Tracking Method Based On Particle Filter Fusion

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17924004

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17924004

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 22/09/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17924004

Country of ref document: EP

Kind code of ref document: A1