Path analysis of tennis ball automatic transmitter based on artificial intelligence technology

Abstract

In order to improve the analysis effect of tennis automatic transmitter path, this study designed a tennis automatic transmitter path analysis method based on artificial intelligence technology. Firstly, the structure of tennis automatic transmitter is designed, including five parts: ball receiving structure, ball sending structure, ball storage structure, launching structure and moving structure. Then, a grid model of the path distribution after the tennis ball is launched is constructed, and the constraint parameters of the path planning of the launcher are optimized. Finally, the force and running track of the tennis ball before and after launching are analyzed to realize the automatic launching path analysis of the tennis ball. The experimental results show that the convergence value of this method is 1.39%, and it only takes 3.5 minutes to generate the path analysis results. Therefore, this method has good application effect in terms of convergence and timeliness, and can be effectively applied to the path analysis of tennis automatic transmitter.

Keywords

Artificial intelligence technology simulation method tennis automatic transmitter path analysis

1. Introduction

Tennis is one of the ball sports, in the middle of the court set up a separate net, the two sides of the game each occupy a side of the court, the players hit the ball with a tennis racket. Tennis originated in France in the 12th and 13th centuries, was born in The United Kingdom, and reached its climax in the United States [1].

At present, tennis has become the world’s second largest gentlemen’s sport. In recent years, with more high-level professional tennis matches held in China and Chinese women tennis players making brilliant achievements in the world, the development of tennis in China has made great progress [2, 3]. As more and more people join in tennis. Tennis sport research also attracted more scientific research workers to participate.

With the development of computer software and hardware technology, using computer technology to analyze and study tennis has attracted the active participation of many domestic researchers. The automation and intellectualization of sports equipment have promoted the vigorous development of tennis in China. Sports and fitness equipment such as automatic transmitter, automatic ball picking machine and man-machine combat have gradually entered People’s Daily life, providing people with new training methods and sports practices [4, 5].

For tennis automatic transmitter, research on its specific launching path is beneficial to improve the quality of the serve and make it more intelligent. To this end, the path tracking method of tennis motion target using infrared spectrometer is designed [6]. The method adopts the sampling module moving targets in infrared spectrometer interference figure recovered spectrum, phase difference comparing with the actual spectrum, building dynamic observation model, and according to the observations using local background weighted mark moving targets, and finally through the local search algorithm for adaptive update dynamic observation model, complete the automatic tracking of moving targets. Experimental results show that this method can keep high tracking success rate under different background, illumination and Angle. The research method of ball rotation flight track of tennis automatic server based on STM32 microcontroller is designed [7]. Aiming at the problem that the traditional method has poor control effect on the ball rotation flight trajectory of tennis automatic server, a new method is studied. By studying the force of tennis ball in the air before being launched and in rotating flight, the conclusion is drawn that force is an important factor affecting the flight trajectory of tennis ball, and the different flight trajectory formed by two kinds of automatic transmitter in the air are compared. 1 000 tennis balls were launched and the number of net passes, effective balls and landing errors were recorded and analyzed. The experimental results show that, compared with the traditional automatic server, STM32 single-chip microcomputer has better control effect of ball rotation flight trajectory, and the number of tennis balls sent by the net is higher, the number of effective balls is more, and the average error of the landing site is smaller, indicating that its performance is stronger. The trajectory dynamics of tennis serve motion is analyzed and simulated [8]. The method holds that the rich practical experience of tennis provides a lot of material for technical analysis of tennis. However, the technical analysis of conventional tennis basically focuses on individual research, and there are few researches on the basic theory of tennis, especially the theoretical analysis on the trajectory of tennis movement. Therefore, the method firstly analyzes the force condition of tennis movement, and finds out the force that affects the trajectory of tennis movement. Then the motion process is described according to the force condition and the motion equation model is established. At last, the simulation trajectories of the two kinds of service technique are simulated, and the key points that should be paid attention to in the practice of tennis service are put forward.

Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a similar way to human intelligence [9]. Research in the field of artificial intelligence includes robotics, language recognition, image recognition, natural language processing and expert systems. Since the birth of artificial intelligence technology, the theory and technology are increasingly mature, and the application field is also expanding. Artificial intelligence is a challenging science, and those who do it must have knowledge of computers, psychology and philosophy. Artificial intelligence is a very broad science, it is composed of different fields, such as machine learning, computer vision and so on. In general, a major goal of ai research is to enable machines to perform complex tasks that normally require human intelligence [10]. Based on the above traditional research, this paper designs the path analysis of tennis automatic transmitter based on artificial intelligence technology.

2. Structure design of tennis automatic transmitter

The overall scheme of tennis automatic transmitter structure designed in this study is shown in Fig. 1.

Figure 1.

Tennis automatic transmitter overall design scheme.

2.1 To collect the ball structure

At present, the most common devices for collecting and picking up balls in the market are manually operated tennis tubes or baskets. Tennis baskets are made of elastic strips of material spaced slightly less than the diameter of a tennis ball that are pressed into the basket when pressed or rolled across the body. The tennis cylinder is composed of collecting cylinder, handle, inlet valve and spring. The handle is easy to operate. The inlet valve makes the tennis ball only enter but not out, and the collecting cylinder is used for storing the ball. These two ways are low in efficiency, low in storage, primitive in operation and have great limitations.

The tennis automatic transmitter designed in this study adopts three blades to collect the ball, and uses the rotation of the blades to collect the ball. In order to improve the efficiency of collecting the ball, the front end of the base is designed as an outer eight-shaped guide rail, which makes it easier to catch the tennis ball. The rear part is designed as an inclined surface, and the power is transferred through a belt drive to rotate the fan blade to propel the ball from the inclined surface into the ball feeding box.

2.2 Send the ball structure

The ball feeding structure of tennis automatic transmitter designed in this study is composed of three parts: ball feeding box, screw and baffle. The bottom of the rear panel of the ball feeding box is provided with a rectangular groove with a height of 80 mm and a length of 240 mm. The bottom plate of the ball feeding box is properly tilted to ensure that the tennis ball can automatically roll to the rectangular groove. A baffle is arranged outside the rectangular slot, through the baffle action to control the tennis ball rolling out at the appropriate time, to avoid the tennis ball rolling out of the rectangular slot in advance in the lifting process of the ball box.

2.3 Store the ball structure

The ball storage structure should be able to accommodate as many tennis balls as possible, the ball frequency is adjustable, not easy to jam the ball, and easy to serve the tennis ball, in order to improve the efficiency of tennis ball launch, reduce the launch time for the design purpose, the design of the device. Ball storage structure is mainly composed of ball storage box, stirring mechanism and ball feeding mechanism, as shown in Fig. 2.

Figure 2.

Ball storage structure drawing of tennis automatic transmitter.

In order to maximize the full use of space, the ball storage box is divided into two layers, the upper layer is designed into the bottom of the concave box structure, the bottom opened a slightly larger than the diameter of the ball hole, to ensure that the tennis ball can smoothly through the hole into the ball feeding mechanism; A rectangular groove is arranged behind the lower layer for placing the ball mechanism. The stirring mechanism is designed as a crank and rocker mechanism, which is mainly used for mild stirring of the tennis balls in the storage box to ensure that the tennis balls can enter the feeding mechanism more reliably. The rod lengths are respectively 70 mm, 145 mm, 180 mm and 200 mm. The connecting rod is designed as a barb shaped rod, and its size is determined according to the curve law of the connecting rod required by the stirring action. The ball feeding mechanism is mainly composed of two parts: the ball feeding plate and the six-minute groove wheel mechanism. Its operation is driven by an electric motor. There are six circular array of ball storage ports on the ball feeding plate. After the tennis ball enters the ball feeding plate through the ball storage box, the six-minute groove wheel mechanism drives the ball feeding plate to rotate under the drive of the motor. When the ball storage port with the tennis ball enters the position aligned with the lower ball feeding port, the tennis ball enters the hose to complete the ball feeding action. Each rotation of the motor, the rotation of the ball plate 60 ${}^{\circ}$ , through the control of driving speed can achieve the control of the ball frequency.

2.4 Launch structure

The launching structure should adapt to various requirements of the service and realize the adjustable speed and direction of the ball. The main body of the launching structure consists of friction wheel mechanism, planetary wheel mechanism and crank and rocker mechanism.

The automatic launching structure adopts the original four-friction wheel service mode. The structure uses four friction wheels for serving, which can produce a larger initial speed of serving. Because the rotation direction of serving is controlled by controlling the speed of four friction wheels, the style of serving is more abundant and the efficiency of serving is higher. The four friction wheels are independently connected with a driving motor, and the space between the four wheels is slightly smaller than the diameter of the tennis ball.

The planetary wheel mechanism controls the left and right steering of the launching mechanism so that the horizontal Angle of the ball can be adjusted. Its structure is shown in Fig. 3.

Figure 3.

Drawing of tennis automatic transmitter.

In Fig. 3, A is the pinion gear, B is the big gear, C is the inner gear, D and E are the linkage shafts, F is the planetary frame, and G is the driving device.

2.5 Mobile structure

The mobile structure is rear-wheel drive, and the front wheel is used to control the steering, similar to the car steering device, by controlling the Angle of the turntable to control the steering of the tennis automatic transmitter. The steering mechanism is a six-bar mechanism with 1 degree of freedom. The Angle of turning required for tennis automatic transmitter is $\pm$ 35 ${}^{\circ}$ $\sim$ $\pm$ 45 ${}^{\circ}$ . According to the maximum steering Angle of the inner and outer wheels of the car, the steering Angle of the tennis automatic transmitter is set as $\pm$ 40 ${}^{\circ}$ . The mechanism can be regarded as a combination of a crank and rocker mechanism and a double crank mechanism, so the motion analysis of the mechanism can be started from the crank and rocker mechanism.

3. Analyze tennis automatic transmitter path based on artificial intelligence technology

3.1 Path distribution grid model of tennis ball after launching

In this paper, the path space of tennis ball is planned and designed by the method of motion primitive path adaptive learning. Firstly, the distribution of position points after the tennis ball was launched was obtained, as shown in Fig. 4.

Figure 4.

The position distribution of the tennis ball after it was launched.

On this basis, the spatial path distributed recombination and optimization control method are adopted to construct the shortest path optimization parameter analysis model of the route planning after the tennis ball is launched. According to the geometric description of the emitter and the environment, the target node optimization of the route planning after the tennis ball is launched is carried out, and the node transmission load is:

$\displaystyle\nabla F(x)=F\cdot\textit{Rot}[v(x)-F]$ (1)

where $F$ represents the objective function of line planning after the tennis ball is launched, and $v(x)$ represents the local optimal parameter set of line distribution after the tennis ball is launched. Then, Newton gradient descent algorithm was used to estimate the intelligent planning parameters of the line after the tennis ball was launched, and the random parameter distribution in the line configuration space was obtained as follows:

$\displaystyle\nabla^{2}F(x)=\frac{1}{\partial F(x)}-2\sum\limits_{j=1}{\nabla F% (x)\times v(x)}$ (2)

where $\partial$ represents the randomness parameter. Combined with the random parameter distribution results, the shortest path optimization parameters from the initial node to the target node are analyzed, and the linear programming model of path parameters is obtained as follows:

$\displaystyle M=l+\frac{1}{\nabla^{2}F(x)}$ (3)

where $l$ represents the translation of the path position of the target node.

In the linear programming model obtained above, based on the analysis of random parameters in the form space on the differential manifold, the route optimization after the tennis ball is launched is conducted, and the path optimization equation thus obtained is:

$\displaystyle L=\int_{0}^{\infty}{M(x)+l}$ (4)

According to the results obtained above, the adaptive spatial parameter design of the line planning after the tennis ball is launched is carried out, and the optimal solution of the line distribution nodes $v_{1}$ , $v_{2}$ and $v_{3}$ can be expressed as:

$\displaystyle f=\frac{({v_{1}+v_{2}+v_{2}})^{2}}{L}$ (5)

The segmented route optimization control method was adopted to design the grid model of line distribution after the tennis ball was launched [11]. Using the shortest path planning and design, the structure model of distributed grid of shortest lines is obtained as shown in Fig. 5.

Figure 5.

The shortest distributed grid structure model of tennis ball after being launched.

In the grid model of path distribution of tennis ball after being launched, artificial intelligence technology is used to analyze the path of tennis ball automatic transmitter.

3.2 Constraint parameter analysis of emitter path planning

The visual servo-oriented mobile route optimization control method is used to locate the nodes of the emitter path and deploy the shortest path optimization. In the node coordinate system, it is assumed that the optimization distribution of constraint parameters of the launcher serving path is $\delta$ . In the configuration space starting from the initial pose of the tennis ball, the estimation model of parameters of the launcher serving path is obtained, which is expressed as follows:

$\displaystyle\mu=\sum\limits_{f-1}{L+\left({\frac{t_{0}-q_{0}}{\delta}}\right)}$ (6)

where $t_{0}$ and $q_{0}$ respectively represent the shortest path distribution distance and pose distance obtained by depth camera [12].

Under visual navigation, fuzzy directional optimization control is carried out for the launcher’s path, and the path distribution fusion control model is established, which is expressed as follows:

$\displaystyle K=p+\sqrt{\eta+\mu}$ (7)

where $\eta$ represents the deep learning fusion feature distribution set of path optimization, and $p$ represents the shortest line fusion parameter. On this basis, combined with the spatial three-dimensional information sampling method, the deep learning gradient function of the serving path of the tennis transmitter is obtained as follows:

$\displaystyle g=\sqrt{\Delta t+\Delta q}+\min K$ (8)

In the formula, $\Delta t$ and $\Delta q$ respectively represent the shortest distribution distance and average pose distance of the launcher’s firing path. Then, three-dimensional model measurement and modeling were adopted, and the characteristic distribution function of the launcher’s path was $z$ [13]. According to each depth projection corresponds to a programmable tennis launching line, the fuzzy control equation of line space planning can be obtained:

$\displaystyle J=\frac{g-z}{e+c}$ (9)

where $e$ represents the path positioning error and $c$ represents the path optimization parameter.

In the process of tennis ball movement, the starting point of path distribution is $P_{0}$ , $V({M_{mi}})=\frac{\sum\limits_{j=1}^{TC}{\textit{is\_visible}({M_{mi},C_{j}})}% }{TC-1}$ , $M\times N$ , $M$ , $N$ , and the normal vector of each point is $i_{0}$ . By adopting the adaptive transmission control method, the analysis model of emitter’s path constraint parameters is as follows:

$\displaystyle\tau=\sum\limits_{1}^{n}{\frac{P_{n}+i_{0}}{J}}$ (10)

3.3 Force analysis of tennis ball before launch

The most important technique in tennis is swinging the tennis racket and hitting the ball. The swinging makes the flight of the tennis ball have certain strength and Angle, while the hitting determines the heaviest flight path and flight distance of the tennis ball, which is the most crucial step for the success of the serving action [14].

In this paper, the launching action of the tennis ball automatic transmitter determines the initial speed, rotation intensity, flight Angle, flight path, flight time and flight distance of the tennis ball in the air. When the tennis ball automatic transmitter is located on the ground, the flight distance of the tennis ball can be calculated by following the formula:

$\displaystyle y=v\times g\times\cos\alpha$ (11)

where $y$ represents the flight distance of the sphere, $v$ represents the flight speed, $g$ represents the acceleration of gravity, and $\alpha$ represents the takeoff Angle of the sphere. When the automatic launcher is at a certain altitude, the altitude $h$ must be imported into the above formula. In this case, the flying distance of the sphere should be calculated as follows:

$\displaystyle y=v\times g\times\cos\alpha\times({\sin\alpha+\cosh})$ (12)

The serve action of tennis automatic transmitter can be divided into three processes:

Process one, when the tennis ball and racket just contact, the tennis ball force elastic deformation, at the same time by the racket reaction force; Process two, when the tennis ball and the racket contact for a period of time, the racket gives the tennis ball the same speed, and the two together in centrifugal motion; In process three, the tennis ball and the racket begin to separate, and the tennis ball obtains the initial speed of flight under the action of the resultant force, so that it can fly in the air [15]. At the moment before the tennis ball flies, the force of the tennis ball is studied, and its specific force is shown in Fig. 6.

Figure 6.

Force analysis of tennis ball before launch.

When the tennis ball and racket just contact, tennis ball in addition to its own gravity, but also by the racket to give it an upward rebound and friction between the air, under the common action of these several forces, the formation of an upward resultant force, so that the tennis ball can rotate in the air flight.

3.4 Force analysis of tennis ball after launch

There are three kinds of paths for serving path. Path Side path spin path and topspin path serve the ball easily. That is, the path of the ball itself is hit by the racket path of the path produced by the path of its own rotation is side path or upspin path. Flat firing the ball of the ball path, path itself is a spiral path transfer path, path is only because the ball is struck path is the main force of the flux path across its path, makes its own path to the path of the ball rotation less path, therefore is called service path to flatten out, however, as the path to hit the ball speed increase the path of the path, the path of the ball speed will improve the path. Because the path of the tennis ball is affected by the air resistance path of the upper path and the lower path is not the same path, the path tennis ball will produce a certain path of rotation path.

Ideally, on a flat serve, the tennis ball doesn’t spin, so we don’t take into account the spin of the tennis ball. In flight, the tennis ball is mainly affected by two forces: gravity straight down and air resistance in the opposite direction of the tennis ball movement. In practice, there will be a certain amount of rotation in the process of tennis movement. When considering the rotation in the process of tennis movement, the tennis ball is not only subjected to gravity and air resistance, but also subjected to Magnus force. The direction of magnus force produced by different rotation directions is also different.

In the process of hitting the ball, tennis side force is basically balanced, the two forces are opposite, and offset each other, but the ball up and down force is not balanced, on the one hand is affected by the ball force (hit the ball, the ball position in the back of the upper part of the ball), on the other hand, the ball movement path is down. Therefore, the upper and lower forces of the ball are different, which is also the cause of Magnus force. Therefore, the force analysis of the ball can only consider the topspin of the ball when the flat shot is served. In this paper, only the topspin of the tennis ball is considered and the force analysis is carried out.

The force of the tennis ball after launching is shown in Fig. 7.

Figure 7.

Force analysis of tennis ball after launch.

The following is a detailed analysis of the three stress conditions:

(1) gravity $F_{g}$ :

$\displaystyle F_{g}=m\times g$ (13)

where $m$ is the mass of the tennis ball and $g$ is the acceleration of gravity. The tennis ball always receives the action of gravity during the movement, the direction is straight down.

(2) Air resistance $F_{d}$ :

$\displaystyle F_{d}=-\frac{1}{2}C_{d}\times\rho\times S\times v$ (14)

where $C_{d}$ represents the resistance coefficient, $\rho$ represents the air density, $S$ represents the windward area of the tennis ball, and $v$ represents the movement speed of the tennis ball. Air resistance is always opposite to the movement of the tennis ball.

(3) Magnus force $F_{m}$

$\displaystyle F_{m}=\frac{\chi}{2}\times\rho\times\frac{d}{2}\times S\times w$ (15)

where $d$ represents the diameter of the tennis ball, $\chi$ represents the lift coefficient, and $w$ represents the angular speed of the tennis ball.

3.5 Simulation method of artificial intelligence technology is used to analyze the path of tennis automatic transmitter after serving

In this section, this study analyzes the path of tennis automatic transmitter after serving by using the simulation method in artificial intelligence technology.

Whether the ball rotates in the air depends on the force exerted by the tennis automatic launcher on the tennis ball. When the tennis automatic launcher hits the tennis ball without passing the center of gravity of the tennis ball, the tennis ball is subjected to uneven force so that the ball rotates. When a racket hits a tennis ball, in addition to acting on the traction of the ball, it is acting on both friction and spring forces. Frictional force and elastic force are the auxiliary forces of traction force. The two forces have the same magnitude of traction force and the opposite effect of gravity. The reverse of the pull is parallel to the ball’s fall line, and friction and spring force combine in one direction, causing the ball to spin.

Among them, the traction helps tennis for linear motion in the direction of the traction, but because of the existence of other force, tennis balls in the air flight path curve, and the size of the size of the spinning force and the traction depends on the size of the lever, thus, tennis from any point of the nature of the flight path is the same as the nature of the pendulum type line.

As the ball rotates, the axis of rotation around it changes constantly. According to the position of the force on the tennis ball, there are three types of spin balls: front spin, right spin, and left spin. The three rotating effects of force applied to different parts of the sphere are shown in Fig. 8.

Figure 8.

Forces act as three rotating effects applied to different parts of the sphere.

In order to better study the flight path of spinning tennis ball and provide more theoretical basis for trainers’ tennis training, this study has carried out in-depth research on the flight path of front spinning ball, left spinning ball and right spinning ball in the air. The flight paths of the three spinning balls are shown in Fig. 9.

Figure 9.

Flight paths of three spinning balls.

The flight path of the forward spinning ball from the tennis ball’s automatic transmitter is a near-perfect parabola, with the ball reaching its highest point of flight directly above the net. Assume that the vertical distance from the automatic server to the net is $D$ . If no one hits the tennis ball, it will land directly opposite the automatic service machine and at the same vertical distance $D$ to the net. Therefore, in order to catch the forward spin ball, the vertical distance between the player and the net should be less than $D$ .

The flight path curve of the left spin ball is similar to that of the right spin ball, but the deviation direction is different. The left spin ball flies in the left direction, and the right spin ball flies in the right direction. The left and right spin ball reach the highest point of flight before flying to the net, and the flight path is a curve. If no one hits the tennis ball, it will fall far away from the automatic tennis machine and at a vertical distance greater than $D$ from the net, so in order to catch the ball spinning left or right, the player needs to move a certain distance to the left or right. On the other hand, by launching different spinning balls, the exercise consumption of the trainer can be increased, and the reaction ability of the trainer can be exercised at the same time, so as to greatly improve the tennis skills of the trainer.

4. Experiment and result analysis

In order to verify the practical application performance of the path analysis method of tennis automatic transmitter designed above based on artificial intelligence technology, the simulation experiment process is designed as follows.

The experimental scene is shown in Fig. 10.

Figure 10.

Experimental scene diagram.

Due to the influence of environmental noise and non-uniform motion of the tennis ball, it is easy to lead to the movement of the tennis ball often accompanied by multiple pixels or even dozens of pixels between the adjacent two frames of the size of the jump, as well as due to the inevitable occlusion in the process of movement, the tennis ball movement target loss. In order to effectively track and locate the tennis target sent by the emitter, the following optimization is made in the process of tracking the tennis moving target: If the detection of the current tennis moving target fails, it indicates that the tennis position of the front and back frames jumps up or the tennis moving target temporarily breaks away from the video image area. To solve this problem, expand the region of interest of tennis target detection, expand the length and width of the region of interest, and find the tennis target in the expanded region of interest.

In order to avoid the uniformity of experimental results, the method presented in this paper was compared with the path tracking method of tennis motion target based on infrared spectrometer in Reference [6] and the ball rotation flight trajectory analysis method of tennis automatic server based on STM32 microcontroller in Reference [7].

First, the convergence ability of different methods is tested, and the comparison results are shown in Fig. 11.

It can be seen from the results in Fig. 11 that the convergence curve of the method in this paper is the lowest and the convergence speed is the fastest. At the 50th iteration, the method in this paper reaches a stable state with a value of 1.47%. Other methods have higher convergence values at this time, 1.85% and 2.26% respectively. Compare the three methods, the convergence performance of the proposed method is better than that of other certification methods, indicating that the proposed method has a strong analysis of automatic tennis transmitter path and can effectively improve the reliability of subsequent tennis path planning results.

On this basis, the time consuming process of path analysis of different methods is tested, and the comparison results are shown in Table 1.

Table 1

Time-consuming comparison of path analysis by different methods (min)

Number of iterations/times	Method of this paper	Method of Reference [6]	Method of Reference [7]
10	2.1	5.0	5.0
20	2.3	5.3	4.7
30	2.5	5.4	5.3
40	2.5	5.9	5.4
50	2.8	6.3	6.0
60	3.2	6.4	6.3
70	3.2	5.9	6.4
80	3.4	7.4	7.0
90	3.5	7.7	8.2
100	3.5	7.8	8.4

Figure 11.

Comparison of convergence of different methods.

By analyzing the results shown in Table 1, it can be seen that the time taken to generate path analysis results by using three different methods is different, but all of them increase with the increase of the number of experiments. Among them, the maximum time for method of Reference [6] to generate path analysis results is 7.8 min, and the maximum time for method of Reference [7] to generate path analysis results is 8.4 min. However, the maximum time for Method of this paper to generate path analysis results was only 3.5 min. In contrast, the path analysis process of method of this paper is shorter, indicating that the Method has higher timeliness.

In conclusion, this study has better application effect in terms of convergence and timeliness, and can be effectively applied to the path analysis of tennis automatic transmitter.

5. Conclusion

Aiming at the problems of poor convergence and time-consuming path analysis of automatic tennis ball launch path, an automatic tennis ball launch path analysis method based on artificial intelligence technology was proposed. In this method, the launcher structure was designed, and the automatic tennis launcher was built. The grid model of the path distribution after the tennis launch was constructed through the adaptive learning of the path of motion primitives, the constraint parameters of the path planning of the launcher were optimized, the motion characteristics were analyzed, and the automatic tennis path analysis was completed. Verification analysis shows that the convergence value of the design method is reduced by 0.38% and 0.79%, respectively, and the analysis time is reduced by 4.3 min and 4.9 min. Therefore, the method can effectively improve the convergence and reduce the time, and has feasibility and effectiveness to solve the above problems. Due to the limitation of research time, this study did not analyze the memory ratio of the method. Subsequent work will further study this aspect in order to reduce the memory ratio and further improve the operation efficiency of the method.

References

Guan

. Tennis motion tracking based on particle filter in Kalman filter prediction. Chin J Electron Devices. 2019; 42(4): 973-977.

Chen

Hou

. Design of vision navigation system for tennis picking robot. J Langfang Teach Univ (Nat Sci Edition). 2020; 20(2): 42-45.

Zhang

. Analysis of the principle of using transcendental instruments in tennis serve technology. Contemp Sports Technol. 2021; 11(20): 29-32.

Sun

Wang

. The mechanism design of a tennis partner machine. Mach Design Manuf Eng. 2020; 49(11): 44-48.

Zhang

Gou

Chen

. The design of an automatic serve device for tennis. Mach. 2020; 47(3): 71-76.

Wang

Yang

Lei

. Moving target tracking method based on infrared spectrometer in tennis match. Nat Sci J Xiangtan Univ. 2018; 40(2): 46-49.

Shi

. Research on sphere rotating flight trajectory of tennis automatic ball machine based on STM32 microcontroller. Autom Instrum. 2020; (10): 125-128.

. Dynamics analysis and simulation research of tennis serving movement trajectory. Comp Simul. 2017; 34(3): 243-246.

Zhou

Meng

. Tennis Hawkeye system based on computer vision and LabVIEW platform. Transducer Microsyst Technol. 2018; 37(7): 102-104+107.

10.

Bao

Zhou

Huo

. On the application of artificial intelligence in tennis technical movement analysis. J Nanjing Sport Inst. 2020; 19(11): 47-51.

11.

Shi

. Comparative kinematical analysis of tennis kick serve with different locations. China Sport Sci Technol. 2020; 56(6): 68-75.

12.

Zhong

Zhang

. Structure design of pitching robot. J Beijing Inf Sci Technol Univ. 2019; 34(4): 93-96.

13.

Pei

. Kalman filtering based best hitting point prediction system of tennis serve. Mod Electron Tech. 2018; 41(11): 162-165+170.

14.

Chen

. Kinect based curve trajectory and speed measurement. Phys Teacher. 2021; 42(10): 50-51.

15.

Han

. Research on the lineshape capture method of tennis forehand error track based on D-P algorithm. J Zhoukou Norm Univ. 2019; 36(5): 86-90.