Experimental device based on table pot robot competition

Abstract

The present invention discloses an experimental device based on a curling robot competition, which belongs to the field of artificial intelligence and automatic control technology, and includes: a curling robot experimental site, a curling robot and a curling referee module, wherein the curling robot experimental site includes a starting area, a throwing area, a free defense area, a scoring area and an autonomous return channel; the curling referee module includes a referee machine, a bracket and a camera, the referee machine is connected to the camera, and the camera is installed on the bracket; the curling robot moves on the curling robot experimental site, the curling referee module is fixed on the curling robot experimental site, and the camera of the curling referee module is fixed to the center of the scoring area through the bracket, and the field of view covers the free defense area and the scoring area. The device designs a curling robot competition experimental site suitable for the field of artificial intelligence, which not only realizes the curling robot intelligently throwing the curling ball and autonomously returning, but also realizes the positioning of the curling ball and the score determination, which can provide important reference value for the development of curling robots.

Description

Technical Field

[0001] The invention relates to the technical field of artificial intelligence and automatic control, and in particular to an experimental device based on a table kettle robot competition. Background Art

[0002] Table curling is a semi-physical simulation competition that reduces the curling venue and curling ball to the table level. Compared with curling, table curling is more conducive to promoting curling among the public, and can reflect the table curling competition strategies formulated by the contestants and their control over the table curling ball. Therefore, there is an urgent need for a table curling robot, a table curling referee module and a suitable competition venue for table curling competitions to realize the table curling robot's intelligent decision-making, intelligent throwing, intelligent return and other full-process autonomous competitions without human participation, provide a reference for the development of curling robots, and lay a technical foundation for conducting human-machine competitions on ice. Summary of the invention

[0003] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the purpose of the present invention is to provide an experimental device based on table kettle robot competition.

[0005] To achieve the above-mentioned purpose, an embodiment of the present invention proposes an experimental device based on a table pot robot competition, including: a table pot experimental site, a table pot robot and a table pot referee module, wherein the table pot experimental site includes a starting area, a throwing area with black and white grids and green straight lines, a free defense area, a scoring area and an autonomous return channel; the table pot robot includes a two-wheel differential chassis, a table pot ball clamp, a visual sensor, an inertial sensor, a motion control card, a core processing board and a battery pack; the table pot referee module includes a referee machine, a bracket and a camera, the referee machine is connected to the camera, and the camera is installed on the bracket; the table pot robot moves on the table pot experimental site, the table pot referee module is fixed on the table pot experimental site, and the camera of the table pot referee module is fixed through the bracket directly above the center of the scoring area, and the field of view covers the free defense area and the scoring area.

[0006] The experimental device based on the curling robot competition in the embodiment of the present invention throws the curling ball to the expected position through the speed control algorithm of the curling robot, and then trains the intelligent algorithm to control the curling robot to return to the initial position; the position of the curling ball in the scoring area is identified by the curling referee module to determine the current score, and the position of the curling ball is sent to the curling robot as the basis for the next throwing strategy, thereby realizing the whole process of autonomous experiment, thereby providing a reference for the development of curling robots and laying a technical foundation for conducting human-machine competitions on ice.

[0007] In addition, the experimental device based on the table kettle robot competition according to the above embodiment of the present invention may also have the following additional technical features:

[0008] Furthermore, in one embodiment of the present invention, the autonomous return channel is a three-quarter elliptical runway, one end of the three-quarter elliptical runway is connected to the starting area, and the other end is connected to the throwing area, and the starting area is connected to the throwing area, the throwing area is connected to the free defense area, and the free defense area is connected to the scoring area, wherein the starting area, the throwing area, the free defense area and the scoring area are on the same horizontal line.

[0009] Furthermore, in one embodiment of the present invention, the core processing board is respectively connected to the visual sensor, the inertial sensor, the motion control card and the table tennis fixture, the motion control card drives the two-wheel differential chassis, and the table tennis fixture is fixed on the two-wheel differential chassis; the battery pack is respectively connected to the two-wheel differential chassis, the visual sensor, the inertial sensor, the motion control card and the core processing board.

[0010] Furthermore, in one embodiment of the present invention, the table pot robot is specifically used for: the visual sensor is used to collect the current image of the table pot experimental field to extract the black and white grids and green straight lines in the throwing area; the core processing board is used to calculate the relative position information of the table pot robot in the table pot experimental field, and use the relative position information as posture feedback data for automatic correction; the motion control card is used to drive the two-wheel differential chassis to move along the green straight line according to the posture feedback data after correction, so as to realize intelligent throwing of the table pot ball.

[0011] Furthermore, in one embodiment of the present invention, the core processing board is also used to train a two-category image classification model to obtain a classification pre-training model, and then use the classification pre-training model to perform neural network operations on the current image to obtain a blocking probability. When the blocking probability is greater than a set value, the motion control card is controlled to drive the two-wheel differential chassis to turn, thereby realizing real-time obstacle avoidance of the table pot robot.

[0012] Furthermore, in one embodiment of the present invention, the visual sensor is also used to provide a human-computer interaction interface to mark the target point position on the current image; the core processing board is also used to train the image regression model to obtain a regression pre-training model, and then use the regression pre-training model to perform neural network operations on the current image to obtain the current target point position, and finally construct a table pot robot motion model to process the current target point position to obtain dual-wheel power, and the motion control card enables the table pot robot to intelligently track on the autonomous return channel according to the dual-wheel power.

[0013] Further, in one embodiment of the present invention, the table curling referee module is specifically used for: the camera is used to collect scene images within its field of view; the referee machine is used to extract four feature points in the scene image, and is used to perform HSV screening and extraction on multiple table curling balls in the scene image to obtain a four-color mask image; a scoring area reference image coordinate system is obtained according to the four feature points and the four-color mask image; two-color table curling balls are screened out in the scoring area reference image coordinate system, morphological opening and closing operations are applied to remove noise points and an area threshold is set to obtain the outline of the table curling ball, the center and radius of the minimum circular boundary of the outline of the table curling ball are obtained, the center is used as the center coordinate of the table curling ball, the score is determined according to the distance from the center coordinate of each two-color table curling ball to the center point of the scoring area, and the center coordinate of each table curling ball is sent to the table curling robot.

[0014] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0016] Figure 1 A schematic diagram of the structure of an experimental device based on a table kettle robot competition according to an embodiment of the present invention;

[0017] Figure 2 A schematic diagram of the structure of a table pot experimental site according to an embodiment of the present invention;

[0018] Figure 3 A schematic diagram of a scoring area structure according to an embodiment of the present invention;

[0019] Figure 4 is a schematic structural diagram of a table pot robot according to an embodiment of the present invention;

[0020] Figure 5 Schematic diagram of the structure of a table tennis referee module according to an embodiment of the present invention.

[0021] Explanation of the reference numerals: 100 - table tennis experimental site, 101 - starting area, 102 - throwing area with black and white grids and green straight lines, 103 - free defense area, 104 - scoring area, 105 - autonomous return channel, 200 - table tennis robot, 201 - two-wheel differential chassis, 202 - table tennis fixture, 203 - visual sensor, 204 - inertial sensor, 205 - motion control card, 206 - core processing board, 207 - battery pack, 300 - table tennis referee module, 301 - referee machine, 302 - bracket and 303 - camera. DETAILED DESCRIPTION

[0022] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0023] The following describes an experimental device based on a table kettle robot competition proposed according to an embodiment of the present invention with reference to the accompanying drawings.

[0024] Figure 1 It is a structural schematic diagram of an experimental device based on a table pot robot competition according to an embodiment of the present invention.

[0025] like Figure 1 As shown, the device 10 includes: a table pot experimental field 100, a table pot robot 200 and a table pot referee module 300. The table pot robot moves on the table pot experimental field, the table pot referee module is fixed on the table pot experimental field, and the camera of the table pot referee module is fixed to the center of the scoring area of ​​the table pot experimental field through a bracket, and the field of view covers the free defense area and the scoring area of ​​the table pot experimental field.

[0026] Furthermore, if Figure 2 and 3 As shown, the table pot experimental site 100 includes a starting area 101, a throwing area 102 with black and white grids and green straight lines, a free defense area 103, a scoring area 104 and an autonomous return channel 105, wherein the autonomous return channel 105 is a three-quarter elliptical runway, one end of the three-quarter elliptical runway is connected to the starting area 101, and the other end is connected to the throwing area 102, and the starting area 101 is connected to the throwing area 102, the throwing area 102 is connected to the free defense area 103, and the free defense area 103 is connected to the scoring area 104, and the starting area 101, the throwing area 102, the free defense area 103 and the scoring area 104 are on a horizontal line.

[0027] Furthermore, if Figure 4As shown, the table pot robot 200 includes a two-wheel differential chassis 201, a table pot ball fixture 202, a visual sensor 203, an inertial sensor 204, a motion control card 205, a core processing board 206 and a battery pack 207, wherein the two-wheel differential chassis 201 is the bottom layer of the robot, the table pot ball fixture 202 is fixed on the two-wheel differential chassis 201, the battery pack 207 is fixed above the two-wheel differential chassis 201, the inertial sensor 204 and the motion control card 205 are fixed above the battery pack 207, and the inertial sensor 204 is arranged on the motion control board 205. On the right side of the card 205, the visual sensor 203 and the core processing board 206 are both installed above the motion control card 205, and the core processing board 206 is placed on the right side of the visual sensor 203. The core processing board 206 is electrically connected to the visual sensor 203, the inertial sensor 204, and the motion control card 205 respectively. The motion control card 205 drives the two-wheel differential chassis 201; the battery pack 207 is respectively connected to the two-wheel differential chassis 201, the visual sensor 203, the inertial sensor 204, the motion control card 205 and the core processing board 206 to provide electrical energy.

[0028] Specifically, the visual sensor 203 is used to collect the current image of the table pot experimental site 100 to extract the black and white grids and the green straight line in the throwing area 102; the core processing board 206 is used to calculate the relative position information of the table pot robot 200 in the table pot experimental site 100, and use the relative position information as posture feedback data for automatic deviation correction; the motion control card 205 is used to drive the two-wheel differential chassis 201 to move along the green straight line according to the posture feedback data after deviation correction, so as to realize the intelligent throwing of the table pot ball;

[0029] The visual sensor 203 is used to collect the current image of the table tennis experimental site 100. The current image may be a scene in which the game has been going on for a period of time and the two sides are confronting each other, that is, one side needs to attack the other side and change its current position; the core processing board 206 is also used to train the two-class image classification model to obtain the classification pre-training model, and then use the classification pre-training model to perform neural network calculations on the current image to obtain the probability of being blocked. When the probability of being blocked is greater than the set value, the motion control card 205 is controlled to drive the two-wheel differential chassis 201 to turn, so as to achieve real-time obstacle avoidance of the table tennis robot 200;

[0030] The visual sensor 203 is also used to collect the current image of the table pot experimental site 100 to provide a human-computer interaction interface to mark the target point position on the current image; the core processing board 206 is used to train the image regression model to obtain a regression pre-training model, and then use the regression pre-training model to perform neural network operations on the current image to obtain the current target point position, and finally construct a table pot robot motion model to process the current target point position and obtain dual-wheel power. The motion control card 205 enables the table pot robot 200 to intelligently track on the autonomous return channel 105 based on the dual-wheel power.

[0031] Furthermore, if Figure 5 As shown, the table tennis referee module 300 includes a referee machine 301, a bracket 302 and a camera 303. The referee machine 301 is connected to the camera 303, and the camera 303 is fixed at a preset height directly above the center of the scoring area 104 through the bracket 302.

[0032] Specifically, the camera 303 is used to collect scene images within its field of view; the referee machine 301 is used to extract four feature points in the scene image, and is used to perform HSV screening and extraction on multiple table tennis balls in the scene image to obtain a four-color mask image; a reference image coordinate system of the scoring area is obtained based on the four feature points and the four-color mask image; a two-color table tennis ball is screened out in the reference image coordinate system of the scoring area, and a morphological opening and closing operation is applied to remove noise points and an area threshold is set to obtain the outline of the table tennis ball, and the center and radius of the minimum circular boundary of the outline of the table tennis ball are obtained, and the center is used as the center coordinate of the table tennis ball, and the score is determined based on the distance from the center coordinate of each two-color table tennis ball to the center point of the scoring area, and the center coordinate of each table tennis ball is sent to the table tennis robot 200.

[0033] Furthermore, the specific working process of the experimental device based on the table kettle robot competition proposed in the embodiment of the present invention is as follows:

[0034] First, the table pot robot 200 moves on the table pot experimental field 100, and the table pot referee module 300 is fixed on the table pot experimental field 200, wherein the camera 303 of the table pot referee module 300 is fixed at a designated position by a bracket 302;

[0035] The speed control algorithm of the table pot robot 200 is then used to throw the table pot ball to the expected position, and then the intelligent algorithm is trained to control the table pot robot to return to the initial position. Specifically, the visual sensor 203 of the table pot robot 200 collects the current image of the table pot experimental site 100, extracts the black and white grids and green straight lines in the throwing area, calculates the relative position of the table pot robot 200 in the table pot experimental site 100, and uses this information as posture feedback data for automatic deviation correction. The motion control card 205 drives the two-wheel differential chassis to move along the green straight line to realize the intelligent throwing function of the table pot ball. At the same time, after the current image is collected, a human-computer interaction interface is provided to mark the target point position on the current image, and the core processing board 206 is used to train the image regression model, and the regression prediction The training model uses a neural network to calculate the current target point position on the currently collected image, and calculates the dual-wheel power according to the motion model of the table pot robot to realize the intelligent tracking of the table pot robot, that is, the target point position is determined according to the current image, and the next throw is made, so that the table pot robot 200 learns intelligent tracking; in addition, after both sides throw the table pot ball many times, multiple ball collisions will occur according to the throwing strategy. Therefore, after the current image is collected, the core processing board 206 is also used to train a two-category image classification model, and the classification pre-training model is deployed to calculate the blocking probability of the currently collected image through a neural network. When the blocking probability is greater than the set value, the motion control card 205 drives the dual-wheel differential chassis to turn, thereby realizing real-time obstacle avoidance of the table pot robot 200.

[0036] Next, the current score is determined by identifying the position of the table pot ball in the scoring area through the table pot referee module, and the position of the table pot ball is sent to the table pot robot 200 as the basis for the next throwing strategy. Specifically, the field of view of the camera 303 of the table pot referee module 300 covers the free defense area 103 and the scoring area 104 (red ring, green ring), and the table pot balls (orange, blue) thrown into the camera field of view are color screened, and the image taken by the camera 303 is transformed into the HSV space, and then the scoring area and the table pot ball are screened according to the HSV screening ranges corresponding to the four colors of red, green, orange and blue, and the mask images are extracted respectively;

[0037] Four feature points are extracted from the image read by the camera 303. The contour with the largest area enclosed by the red mask image is selected through color screening, and then the rectangular boundary of the contour is obtained; the neighborhood of the four sides of the circumscribed rectangle of the red ring in the green mask image is searched, and the green pixel closest to the center of the red ring is found as the feature point. The coordinates of the four feature points can be obtained in the four directions of up, down, left, and right, such as Figure 3As shown, the known coordinates of the corresponding four feature points in the table pot experimental site design diagram and the four feature point coordinates obtained from the image read by the camera 303 are combined into point pairs, the transmission transformation matrix is ​​calculated, and the image read by the camera is subjected to transmission transformation based on the matrix, and the image is transformed into the reference image coordinate system;

[0038] After the image read by the camera 303 is transformed into the reference image coordinate system of the scoring area, the orange and blue table tennis balls are extracted by color screening, and the morphological opening and closing operations are applied to remove noise and set the area threshold to obtain the outline of the table tennis ball. The center and radius of the minimum circular boundary of the outline are obtained, and the center is used as the center coordinate of the table tennis ball. The score is determined according to the distance from the center coordinate of each orange and blue table tennis ball to the center point of the scoring area (the coordinates are known), and the center coordinate of each table tennis ball is sent to the table tennis robot 200 via the wireless network, and the table tennis robot 200 can then determine the throwing strategy based on the center coordinate.

[0039] In summary, according to the experimental device based on the table curling robot competition proposed in the embodiment of the present invention, the table curling ball is thrown to the expected position through the speed control algorithm of the table curling robot, and then the intelligent algorithm is trained to control the table curling robot to return to the initial position; the position of the table curling ball in the scoring area is identified by the table curling referee module to determine the current score, that is, the intelligent throwing and detection of the table curling ball can be realized, and the position of the table curling ball is sent to the table curling robot as the basis for the next throwing strategy, realizing the whole process of autonomous experiment, which can have important reference value for the development of curling robots and lay a technical foundation for conducting human-machine confrontation competitions on ice.

[0040] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0041] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0042] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. An experimental device based on a table pot robot competition, comprising: Table pot experimental site, table pot robot and table pot referee module, including: The table pot experiment site includes a starting area, a throwing area with black and white grids and green straight lines, a free defense area, and a scoring area; The table pot robot includes a two-wheel differential chassis, a table pot ball fixture, a visual sensor, an inertial sensor, a motion control card, a core processing board and a battery pack; wherein the two-wheel differential chassis is the bottom layer of the robot, the table pot ball fixture is fixed on the two-wheel differential chassis, the battery pack is fixed above the two-wheel differential chassis, the inertial sensor and the motion control card are fixed above the battery pack, and the inertial sensor is arranged on the right side of the motion control card, the visual sensor and the core processing board are both installed above the motion control card, and the core processing board is placed on the right side of the visual sensor, the core processing board is electrically connected to the visual sensor, the inertial sensor, and the motion control card, and the motion control card drives the two-wheel differential chassis; the battery pack is respectively connected to the two-wheel differential chassis, the visual sensor, the inertial sensor, the motion control card and the core processing board to provide electrical energy; The table pot referee module includes a bracket and a camera, and the camera is installed on the bracket; The table pot robot moves on the table pot experimental field, the table pot referee module is fixed on the table pot experimental field, and the camera of the table pot referee module is fixed through the bracket just above the center of the scoring area, and the field of view covers the free defense area and the scoring area; Characterized in that the table pot experimental site also includes an autonomous return channel, The autonomous return channel is a three-quarter elliptical runway, one end of the three-quarter elliptical runway is connected to the starting area, and the other end is connected to the throwing area, and the starting area is connected to the throwing area, the throwing area is connected to the free defense area, and the free defense area is connected to the scoring area, wherein the starting area, the throwing area, the free defense area and the scoring area are on a horizontal line; The table pot referee module also includes a referee machine, which is connected to the camera, and the camera is used to collect scene images within its field of view; The referee machine is used to extract four feature points in the scene image, and to perform HSV screening and extraction on multiple table tennis balls in the scene image to obtain a four-color mask image; obtain a scoring area reference image coordinate system based on the four feature points and the four-color mask image; screen out two-color table tennis balls in the scoring area reference image coordinate system, apply morphological opening and closing operations to remove noise points and set an area threshold to obtain the table tennis ball outline, obtain the center and radius of the minimum circular boundary of the table tennis ball outline, use the center as the center coordinate of the table tennis ball, determine the score based on the distance from the center coordinate of each two-color table tennis ball to the center point of the scoring area, and send the center coordinate of each table tennis ball to the table tennis robot; specifically, the camera field of view of the table tennis referee module covers the free defense area and the scoring area, performs color screening on the table tennis balls thrown into the camera field of view, transforms the image taken by the camera into the HSV space, and then screens the scoring area and the table tennis balls according to the HSV screening ranges corresponding to the four colors of red, green, orange and blue, and extracts the mask image respectively; Extract four feature points from the image read by the camera, select the contour with the largest area enclosed by the red mask image through color screening, and then obtain the rectangular boundary of the contour; search the neighborhood of the four sides of the circumscribed rectangle of the red ring in the green mask image, find the green pixel closest to the center of the red ring and take it as the feature point, and the coordinates of the four feature points can be obtained in the four directions of up, down, left and right. Combine the known coordinates of the corresponding four feature points in the design diagram of the table pot experimental site and the coordinates of the four feature points obtained from the image read by the camera into point pairs, calculate the transmission transformation matrix, and perform transmission transformation on the image read by the camera based on the matrix, and transform the image into the reference image coordinate system; After transforming the image read by the camera into the reference image coordinate system of the scoring area, the orange and blue table tennis balls are extracted through color screening, and the morphological opening and closing operations are applied to remove noise and set the area threshold to obtain the outline of the table tennis ball. The center and radius of the minimum circular boundary of the outline are obtained, and the center is used as the center coordinate of the table tennis ball. The score is determined according to the distance from the center coordinate of each orange and blue table tennis ball to the center point of the scoring area, and the center coordinate of each table tennis ball is sent to the table tennis robot through the wireless network, and the table tennis robot determines the throwing strategy according to the center coordinate.

2. The experimental device based on the table pot robot competition according to claim 1 is characterized in that: The core processing board is respectively connected to the visual sensor, the inertial sensor, and the motion control card, the motion control card drives the two-wheel differential chassis, and the table tennis fixture is fixed on the two-wheel differential chassis; The battery pack is respectively connected to the dual-wheel differential chassis, the visual sensor, the inertial sensor, the motion control card and the core processing board.

3. The experimental device based on the table pot robot competition according to claim 1 is characterized in that: The table pot robot is specifically used for: The visual sensor is used to collect the current image of the table pot experimental site to extract the black and white grids and green straight lines in the throwing area; The core processing board is used to calculate the relative position information of the table pot robot in the table pot experimental field, and use the relative position information as posture feedback data for automatic deviation correction; The motion control card is used to drive the two-wheel differential chassis to move along the green straight line according to the posture feedback data after correction, so as to realize the intelligent throwing of the table tennis ball.

4. The experimental device based on the table pot robot competition according to claim 3 is characterized in that: The core processing board is also used to train two-category image classification models to obtain a classification pre-training model, and then use the classification pre-training model to perform neural network operations on the current image to obtain the probability of being blocked. When the probability of being blocked is greater than a set value, the motion control card is controlled to drive the two-wheel differential chassis to turn, thereby realizing real-time obstacle avoidance of the table pot robot.

5. The experimental device based on the table kettle robot competition according to claim 3 is characterized in that The visual sensor is also used to provide a human-computer interaction interface to mark the target point position on the current image; The core processing board is also used to train the image regression model to obtain a regression pre-training model, and then use the regression pre-training model to perform neural network operations on the current image to obtain the current target point position, and finally construct a table pot robot motion model to process the current target point position to obtain dual-wheel power. The motion control card enables the table pot robot to intelligently track on the autonomous return channel based on the dual-wheel power.

Images