CN112007343A

CN112007343A - Double-arm boxing training robot

Info

Publication number: CN112007343A
Application number: CN202010649338.5A
Authority: CN
Inventors: 林剑楚
Original assignee: Huaiyin Institute of Technology
Current assignee: Huaiyin Institute of Technology
Priority date: 2020-07-08
Filing date: 2020-07-08
Publication date: 2020-12-01

Abstract

The invention relates to a double-arm boxing training robot. The robot comprises a stereoscopic vision device, a target area device, a mechanical double-arm device, a data processing device and an information display interaction device; the stereoscopic vision device is used for providing measurement of motion data of the boxer; the target area device is used for providing attack training of athletes; the mechanical double-arm device is used for simulating the boxing attack of a real person and providing defense training for athletes; the data processing device is used for reading and analyzing pressure sensor data and stereoscopic vision image data by software and operating and controlling the double mechanical arms; the information display interaction device provides functions of scoring the motion data and the motion postures, displaying error prompts and performing man-machine interaction. The boxing training device can relieve the public demands on coaches of boxing sports, improve the standardization level of boxing teaching, provide professional boxing training and enhance the entertainment of boxing sports.

Description

Double-arm boxing training robot

Technical Field

The invention relates to the field of application of a computer technology auxiliary robot to sports, in particular to a double-arm boxing training robot.

Background

Nowadays, China also carries out quality education while paying attention to cultural education, strengthens the development of physical activities, and actively promotes boxing. The State general office of sports, the sports administration of the State Living, Liqiang, means "the boxing movement into the world college student's sports association would be a necessary trend". At present, nearly 3000 colleges exist in China, more than 30 sports colleges such as Beijing sports college, Shanghai sports college and Wuhan sports college have opened boxing professional courses, and other colleges and secondary schools are increasing boxing class repair courses continuously. The number of 660 cities in China is about, numerous sports and fitness places are provided in the cities, and many sports and fitness places are formed into boxing clubs or communities and gather a large number of boxing enthusiasts. The boxing training aid is not only beneficial to deep popularization of boxing sports, but also beneficial to improving physical and psychological qualities of citizens and students and enhancing self-protection capability, especially for women. However, according to the statistics of the well-known professional boxing data website Boxrec, the number of professional boxers in China is only 498 (rank 7), the number of professional boxers is only 2, and certain gaps are left between the number of professional boxers and other countries. From this it can be calculated that on average there are less than 1 professional boxer per city and not 1/6 per college. In conclusion, with the deep popularization of boxing sports, China will meet the requirements of a large number of boxing coaches and the situation that the number of professionals with actual combat experience is seriously insufficient.

With the continuous maturity of the robot technology, the research on boxing robots at home and abroad is gradually becoming a hot spot. These studies show that boxing training robots can be divided into two major directions of research: 1) a mechanical arm-free type; 2) there are arm types.

For the type without mechanical arms, a boxing teaching robot based on a pressure sensor and an embedded system is designed in the early years of 2017. When a boxer strikes an air bag in the sandbag, the robot can display the force of the boxer attack in real time, and the man-machine interaction is enhanced through functions such as integrated voice broadcasting, however, due to the fact that only the sandbag mode is adopted for training, although the robot completes the function of measuring the force of athletes, the robot cannot realize the function of imitating humanized attack. 2018, German scientists successfully developed a BotBox partner robot. The robot is similar to a rocker in shape and can swing around 360 degrees. A cylindrical boxing target (or a sand bag) is fixed on the upper end part of the robot, and meanwhile, a high-speed computer vision and action recognition system is also arranged, so that the attack of a boxer can be quickly avoided, and the function of accompanying training is realized. The robot is trained in a sand bag mode, functions such as dodging and entertainment are provided, but the sand bag cannot be used for boxing to assault and train the defense capability of athletes, and the problem of insufficient humanoid performance exists. In 2019, Huangxin and the like design and research desktop type boxing robots, the robots have the advantages of small occupied space, low noise and high sensitivity, the shape design of the robots is similar to that of BotBox, a cylindrical boxing sandbag is arranged at the upper end of the robots, and the bases of the robots are made of rubber materials and are damped in a spring mode, so that the robots are easy to serve as small entertainment facilities indoors. The research shows that: the armless boxing robot can effectively finish the attack training of athletes, but is difficult to finish the defense training of the athletes, and has the problem of insufficient accompany training.

Compared with a humanoid robot with a mechanical arm type, especially a double-arm structure, the humanoid robot can better simulate a human being and improve the sense of reality of training. Li Yong soldiers and the like in 2019 realize a family entertainment boxing robot based on mobile phone WeChat remote control, the rolling speed of a robot chassis is remotely controlled by designing a fuzzy control algorithm, and the remote control robot finishes simple boxing actions by two arms. This study provides a practical method of accurately controlling the movement of a home boxing amusement machine. The above studies show that: the existing boxing robot with the double-arm structure has the advantages of human-like performance, capability of improving man-machine interaction and entertainment, and lack of the function of specialized boxing training.

In conclusion, the armless robot has the defect of insufficient anthropomorphic simulation, and the double-arm robot with anthropomorphic simulation cannot realize professional boxing training, so the invention of the boxing training robot with the double-arm structure can effectively simulate real boxing action, improve defense training of athletes and overcome the defect of insufficient anthropomorphic simulation of the existing design.

Disclosure of Invention

In view of the above problems, the present invention provides a two-arm boxing training robot, which is based on computer vision and robot technology, and can provide not only the attacking training of boxing sports, but also the defense training of boxing sports, thereby solving the problem of mass demands of the public on boxing training, and simultaneously enhancing the entertainment interactivity of boxing sports.

The technical scheme of the invention is as follows: a double-arm boxing and training robot, which comprises a stereoscopic vision device, a mechanical double-arm device, a target area device, a data processing device and an information display interaction device which are connected with each other,

the stereoscopic vision device comprises a depth camera and a camera bracket,

the mechanical double-arm device comprises a robot arm imitating a human arm, an imitated fist, a fist cover and a trunk support,

the target area device comprises a sand bag, a stabilizing bracket and a pressure sensor,

the data processing device comprises a stereoscopic vision data reading and analyzing module, a pressure sensor data reading and analyzing module, a robot double-arm control module, a virtual model data processing module and a data storage module,

the information display interaction device comprises an information display module and an information interaction operation module.

Further, the depth camera is supported by the camera bracket and is arranged on the upper side or the lower side of the sandbag; the camera bracket is fixedly arranged at the rear part, the upper part or the left and right sides of the trunk bracket;

the mechanical arms of the humanoid arms are respectively and fixedly arranged at two sides of the trunk support, and the heights of the mechanical arms of the humanoid arms at the two sides are consistent; the simulated fist is respectively arranged on the mechanical arms of the simulated arms at two sides and at one end far away from the trunk support; the boxing glove covers the outer part of the simulated fist;

the sandbag is fixed by the stabilizing support and is arranged above the trunk support; the pressure sensor is arranged inside the sand bag;

the stabilizing support is fixed behind or above the trunk support.

Furthermore, the depth camera, the robot arm of the humanoid arm, the pressure sensor and the information display interaction device are respectively connected to the data processing device.

Further, the sandbags in the target area device and the torso support in the mechanical double-arm device are bilaterally symmetrical about the same plane perpendicular to the ground.

Further, the distance range from the depth camera to the center of the sand bag is 0.1-4 meters;

the lens plane of the depth camera forms a dihedral angle with a plane perpendicular to the ground where connecting lines of fixed midpoints of mechanical arms of humanoid arms on two sides of the trunk support are located, the included angle ranges from 0 degree to 90 degrees and is changed from the lens plane to the plane perpendicular to the ground.

Further, the degree of freedom of the robot arm imitating the human arm is between 5 and 7; the length of the mechanical arm of the humanoid arm is between 20cm and 120 cm.

Further, the height of the sand bag is 10cm to 50cm, and the diameter width of the sand bag is 10cm to 50 cm.

The invention has the beneficial effects that: the double-arm boxing robot can provide attack training for hitting sandbags and defense training for defending mechanical arm attacks, can play a role in correcting and guiding action postures of athletes, provides standard boxing teaching and training, can increase entertainment of the sports, can greatly reduce the requirements on boxing coaches, and promotes the sports and the body health of citizens.

Drawings

FIG. 1 is a schematic diagram of the structure and organization of the present invention;

FIG. 2 is a schematic view of an embodiment of the present invention;

FIG. 3 is a schematic representation of a human skeletal model of the present invention;

FIG. 4 is a schematic diagram of a topological warped mesh of a 3D character model according to the present invention;

in the figure, 1 is a stereoscopic vision device, 11 is a depth camera, and 12 is a camera support;

2 is a mechanical double-arm device, 21 is a robot arm imitating a human arm, 22 is an artificial fist, 23 is a fist cover, and 24 is a trunk support;

3 is the target area device, 31 is the sandbag, 32 is the stabilizing stent, 33 is the pressure sensor;

4 is a data processing device, 41 is a stereoscopic vision data reading and analyzing module, 42 is a pressure sensor data reading and analyzing module, 43 is a robot double-arm control module, 44 is a virtual model data processing module, 45 is a data storage module;

5 is an information display interaction device, 51 is an information display module, and 52 is an information interaction operation module.

Detailed Description

In order to more clearly illustrate the technical solution of the present invention, the following detailed description is made with reference to the accompanying drawings:

as shown in the figure, the double-arm boxing training robot comprises a stereoscopic vision device 1, a mechanical double-arm device 2, a target area device 3, a data processing device 4 and an information display interaction device 5 which are connected with each other,

the stereo vision device 1 comprises a depth camera 11 and a camera support 12,

the mechanical double-arm device 2 comprises a robot arm 21 imitating a human arm, an imitated fist 22, a fist cover 23 and a trunk support 24,

the target area device 3 comprises a sand bag 31, a stabilizing bracket 32 and a pressure sensor 33,

the data processing device 4 comprises a stereoscopic vision data reading and analyzing module 41, a pressure sensor data reading and analyzing module 42, a robot double-arm control module 43, a virtual model data processing module 44 and a data storage module 45,

the information display interaction device 5 comprises an information display module 51 and an information interaction operation module 52;

wherein, normally, the target area device 3 is the head of the partner robot, the stereoscopic vision device 1 is the vision system of the partner robot, the depth camera 11 in the stereoscopic vision device 1 is arranged on the upper side of the sand bag 31 in the target area device 3 through the camera bracket 12;

the data processing device 4 is connected with the partner training robot through a wire line;

the information display interaction device 5 is connected with the partner training robot through a wired line.

Further, the depth camera 11 is arranged on the upper side or the lower side of the sand bag 31 through a camera bracket support 12; the camera bracket 12 is fixedly arranged at the rear, above or any one of the left side and the right side of the trunk bracket 24, namely supported in a floor mode or a suspension mode;

the mechanical arms 21 imitating the human arms are respectively and fixedly arranged at two sides of the trunk support 24, and the heights of the mechanical arms 21 imitating the human arms at the two sides are consistent; the simulated fist 22 is respectively arranged on the mechanical arms 21 of the simulated human arms at two sides and one end far away from the trunk support 24; the boxing glove 23 covers the external part of the simulated fist 22;

the sandbag 31 is fixed by a stabilizing bracket 32 and is arranged on the upper side of the trunk bracket 24; the pressure sensor 33 is arranged inside the sandbag 31;

the stabilizing support 32 is fixed to the rear or above the torso support 24.

Further, the stereoscopic vision device 1, the mechanical double-arm device 2, the target area device 3 and the information display interaction device 5 are respectively connected with the data processing device 4; specifically, the depth camera 11, the robot arm 21 imitating a human arm, the pressure sensor 33 and the information display interaction device 5 are respectively connected to the data processing device 4;

specifically, the stereoscopic vision device 1 is connected to a stereoscopic vision data reading and analyzing module 41 in the data processing device 4 through a depth camera 11 mounted on a camera support 12; the mechanical double-arm device 2 is respectively connected with a robot double-arm control module 43 in the data processing device 4 through a robot arm 21 imitating a human arm; the target area device 3 is connected with a pressure sensor data reading and analyzing module 42 in the data processing device 4 through a pressure sensor 33; the information display interaction device 5 is connected to the virtual model data processing module 44 in the data processing device 4;

the stereoscopic vision device 1 collects movement information of the athlete in the lens, and then a fist detection and human body action posture extraction submodule in a stereoscopic vision data reading and analyzing module 41 in the data processing device 4 is used for detecting a fist target and positioning key points of the body of the athlete in a machine learning mode, and the key points are combined with depth information of a stereoscopic vision camera to determine when spatial position information is calculated;

the stereoscopic vision apparatus 1, which uses a camera with Depth information, such as a PicoZense Depth camera 11 (with a 480 × 640 frame rate of 30FPS or 320 × 480 frame rate of 60FPS) placed above or behind and above a sandbag 31 in the target apparatus 3, captures image RGB data and Depth data of a boxer, and transmits the RGB-D data to the data processing apparatus 4, provides motion information of the boxer: punching speed, human motion posture and spatial position.

The depth camera 11 can provide image RGB data in the field of view of the camera and the depth or distance D information of each image number area, so as to record the movement information of athletes.

The camera bracket 12 is formed by combining a plurality of sub-brackets, the depth camera 11 is fixed at the top of the camera bracket 12, and the height of the position of the fixed point of the depth camera 11 in the camera bracket 12 can be adjusted up and down; preferably, the depth camera 11 has a height ranging from 180cm to 220 cm.

Furthermore, the mechanical double-arm device 2 is composed of two humanoid arm robotic arms 21, preferably, the single humanoid arm robotic arm 21 has a length ranging from 20cm to 120cm, the double arms are fixed on the trunk support 24, the height of the double arms can be adjusted, and the range is 110cm to 170 cm;

the robot arm 21 of the humanoid arm is controlled by 5 to 7 mechanical degrees of freedom to form the rotation function of a shoulder, an elbow and a wrist in the robot arm so as to achieve the purpose of simulating the function of the robot arm, and preferably, the robot arm degree of freedom is 3 degrees of freedom of the shoulder, 2-3 degrees of freedom of a forearm and 0-1 degree of freedom of the wrist, as shown in figure 2;

the degree of freedom of the mechanical double-arm device 2 is controlled by a robot double-arm control module 43 in the data processing device 4, and a basic defense training function is provided;

the rotation information of the degree of freedom in each robot double arm is controlled and updated, so that the attacking action of a boxing partner is simulated, and defense training such as basic straight boxing, hook boxing, punch boxing, combined boxing and the like is provided;

emulation fist 22 connects its one end of keeping away from trunk support 24 on the robotic arm 21 of imitative people's arm, and the boxing glove 23 of boxing is put on emulation fist 22's outside, and the real boxer of simulation has buffering decompression material in emulation fist 22 and the boxing glove 23 simultaneously, protection sportsman's safety.

The sandbags 31 of the target area device 3 are arranged on the stabilizing bracket 32 so as to achieve the purpose of reducing the influence of vibration generated during attack on the working stability of other devices; the size of the sand bag 31 is equivalent to that of the head of a person, the sand bag is cylindrical or ellipsoidal, the height of the sand bag 31 is adjustable, preferably, the height of the sand bag 31 is 10cm to 50cm, the diameter width of the sand bag is 10cm to 50cm, and a pressure sensor 33 is arranged in the sand bag;

the pressure sensor 33 is used for measuring the attack force of the athlete when the athlete attacks the sandbag 31 and transmitting the attack force data to the pressure sensor data reading and analyzing module 42 in the data processing device 4;

the stereoscopic vision data reading and analyzing module 41 in the data processing device 4 obtains the action postures of the athletes acquired by the depth camera 11 in the stereoscopic vision device 1 by utilizing the human body action posture extraction submodule;

acquiring key points of main action postures comprises the following steps: 1 key point of each of the head, the neck and the waist, and 2 key points of each of the shoulder, the elbow, the wrist, the hip bone, the knee and the ankle;

further, the stereoscopic vision data reading and analyzing module 41 in the data processing device 4 mainly includes: a human body action posture extraction submodule, a fist part target detection submodule and an action posture analysis and evaluation submodule;

in the data processing device 4, a human action posture extraction submodule in the stereoscopic vision data reading and analyzing module 41 extracts human key points and generates a human skeleton model as shown in fig. 3;

the fist target detection submodule in the stereoscopic vision data reading and analyzing module 41 in the data processing device 4 detects the moving speed of left and right fists of the athlete in real time, stores the obtained speed result in the data storage module 45, and then displays the numerical value through the information display module 51 in the information display interaction device 5;

the motion posture analysis and evaluation submodule in the stereoscopic vision data reading and analysis module 41 in the data processing device 4 evaluates the correctness and the accuracy of the motion posture of the athlete, and the evaluation result is displayed on the information display module 51 in the information display interaction device 5;

the virtual model data processing module 44 in the data processing device 4 generates a corresponding action posture by combining the 3D virtual athlete model in the data storage module 45 according to the athlete action posture information extracted by the stereoscopic vision data reading and analyzing module 41, and displays the 3D virtual athlete of the generated action posture on the information display module 51 in the information display interaction device 5;

the virtual model data processing module 44 in the data processing device 4 implements the action posture deformation of the 3D character model by using a 3D image topological deformation technology or a human skeleton model technology, and the topological deformed mesh of the 3D character model is as shown in fig. 4;

further, when the athlete selects a certain type or a certain combination type of defense training for a certain duration, the robot double-arm control module 43 in the data processing device 4 provides corresponding defense training by reading defense training data (a freedom control sequence of the robot arm 21 of the humanoid arm) stored in the data storage module 45 in the data processing device 4;

further, the data storage module 45 in the data processing device 4 stores stereoscopic vision RGB-D data, extracted motion attitude information data, an analysis evaluation result, pressure sensor data, interactive data of an interactive operation module, 3D virtual model data, and defense training data in the stereoscopic vision data reading and analyzing module 41;

further, the information display module 51 in the information display interaction device 5 mainly includes hardware such as a display screen; the information interaction operation module 52 mainly includes hardware such as touch devices;

the information display module 51 in the information display interaction device 5 displays the 3D virtual athlete of the generated action posture, and simultaneously displays a fist speed value, an attack force value, an action evaluation score, a current action posture error position, and a prompt of the error position and the content of defense training;

the information interaction operation module 52 in the information display interaction device 5 displays different directions of the 3D virtual athlete according to the interaction operation, so as to provide all-around observation. In addition, the device provides a man-machine interaction function for defense training selection, and the double-arm part can provide defense training of different types and different durations;

in addition, the devices are connected with each other through a wire line or through wireless signals, wherein the three parts of the stereoscopic vision device 1, the mechanical double-arm device 2 and the target area device 3 are mutually independent in space;

the stereoscopic vision apparatus 1 transmits the captured depth and RGB data to the data processing apparatus 4;

the mechanical double-arm device 2 is controlled by the data processing device 4 to make human-like motion;

the target area device 3 measures the attack force of the fist part and transmits the data to the data processing device 4;

the data processing device 4 controls the content displayed on the information display module 51;

the information display module 51 displays the related information controlled by the stereoscopic vision device 1, the mechanical double-arm device 2 and the target area device 3;

the stereoscopic vision device 1 captures images RGB data and depth D data of a boxing player and is used for providing motion information of the player: fist speed, human motion posture and spatial position;

the sandbag 31, fixed by a stabilizing bracket 32, is arranged above the trunk bracket 24; the independent function of the stabilizing support 24 is to achieve the purpose of reducing the influence of the vibration generated during the attack on the working stability of other devices;

further, the sandbag 31 in the target device 3 and the torso support 24 in the mechanical double-arm device 2 are bilaterally symmetrical about the same plane perpendicular to the ground.

Further, the distance between the depth camera 11 and the center of the sand bag 31 ranges from 0.1 meter to 4 meters;

the lens plane of the depth camera 11 forms a dihedral angle with a plane perpendicular to the ground where the connecting line of the fixed midpoints of the robot arms 21 of the humanoid arms on the two sides of the trunk support 24 is located, and the included angle ranges from 0 degree to 90 degrees from the lens plane to the plane perpendicular to the ground.

Further, the degree of freedom of the humanoid arm robot arm 21 is between 5 and 7;

the length of the robot arm 21 of the humanoid arm is between 20cm and 120 cm.

Further, the height of the sand bag 31 is 10cm to 50cm, and the diameter width thereof is 10cm to 50 cm.

The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed.

Claims

1. A double-arm boxing and training robot is characterized by comprising a stereoscopic vision device, a mechanical double-arm device, a target area device, a data processing device and an information display interaction device which are connected with each other,

the stereoscopic vision device comprises a depth camera and a camera bracket,

2. A double-arm boxing training robot as claimed in claim 1, wherein the depth camera is supported by a camera support and is installed on the upper or lower side of the sandbag; the camera bracket is fixedly arranged at the rear part, the upper part or the left and right sides of the trunk bracket;

the sandbag is fixed by the stabilizing support and is arranged on the upper side of the trunk support; the pressure sensor is arranged inside the sand bag;

the stabilizing support is fixed behind or above the trunk support.

3. A double-arm boxing and training robot as claimed in claims 1 and 2, wherein the depth camera, the humanoid arm robot arm, the pressure sensor and the information display interaction device are respectively connected with the data processing device.

4. A two-armed boxing and training robot as claimed in claims 1 and 2, wherein the sandbags in the target area device and the torso support in the mechanical two-arm device are bilaterally symmetric about the same plane perpendicular to the ground.

5. A two-armed boxing training robot as claimed in claim 2, wherein the distance from the depth camera to the center of the sandbag is in the range of 0.1 to 4 meters;

the lens plane of the depth camera, a plane perpendicular to the ground where the connecting line of the fixed midpoints of the mechanical arms of the humanoid arms on the two sides of the lens plane of the depth camera is located form a dihedral angle, the dihedral angle is changed from the lens plane to the plane perpendicular to the ground, and the included angle range is 0-90 degrees.

6. A boxer-sparing robot of the two-armed type according to claim 1, wherein the degree of freedom of the arm of the humanoid arm is between 5 and 7; the length of the mechanical arm of the humanoid arm is between 20cm and 120 cm.

7. A boxer-sparing robot of the two-arm type according to claim 1 wherein the sandbag has a height of 10cm to 50cm and a diameter of 10cm to 50cm in width.