Robotic arm integrated with an elastic tube liquid transfer mechanism and a continuous robot

Abstract

The application discloses a robot integrating an elastic tube liquid transmission mechanism and a continuous robot elastic arm, and relates to the field of robots.The robot comprises a rack (1), an elastic tube (2), support discs, base discs and a driving line (3), the elastic tube (2) is fixed on the rack (1), the elastic tube (2) is divided into multiple arm sections, each arm section is sequentially provided with a base disc and multiple support discs, the interval distance of each support disc is equal, the base disc and the support disc are fixed on the elastic tube (2) through a clamp, multiple through holes are distributed on the support disc, and every three through holes with an interval of 120 DEG form a group.The elastic tube liquid transmission mechanism and the continuous robot elastic arm are integrally designed, an elastic support body does not need to be additionally provided, the robot structure is more compact, the manufacturing cost is lower, and the problem that the prior art cannot well correct the pose error in a robot kinematics solving process is solved.

Description

Technical Field

[0001] This invention relates to the field of robotics, and in particular to a robot that integrates an elastic tube liquid transport mechanism with a continuous robotic elastic arm. Background Technology

[0002] With the development of technology, robots have been widely used in production and daily life. Compared with traditional rigid multi-joint robots, flexible robots have attracted increasing attention and research due to their superior adaptability to complex environments, high interaction safety, and good obstacle avoidance capabilities. Common types of flexible robots include pneumatic actuation, linear actuation, and SMA actuation. Flexible robots have broad application prospects. However, the precise motion trajectory control of flexible robots still faces many challenges. Common methods for describing the deformation of flexible robots include constant curvature methods, cosserat beams, and the finite element method. Piecewise constant curvature, with its simplicity and intuitiveness, is widely used in kinematic modeling.

[0003] However, existing flexible robotic arms require the installation of corresponding auxiliary mechanisms when performing operations, resulting in complex structures. Furthermore, existing technologies cannot effectively correct pose errors during the robot's kinematics solution process. Summary of the Invention

[0004] To address the aforementioned shortcomings in the prior art, the present invention provides a robot that integrates an elastic tube liquid transport mechanism with a continuous robot elastic arm, solving the problems of complex structure and inability to effectively correct pose errors in the robot's kinematics solution process in existing flexible robotic arms.

[0005] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: a robot integrating an elastic tube liquid transfer mechanism with a continuous robot elastic arm, comprising a frame, an elastic tube, a support disk, a base disk, and a drive line. The elastic tube is fixed to the frame and is divided into multiple arm segments. Each arm segment is sequentially provided with a base disk and multiple support disks, and the spacing between each support disk is equal. The base disk and the support disks are both fixed to the elastic tube by clamps. The support disks are provided with multiple through holes, and every three through holes spaced at 120° intervals form a group for the drive line of one arm segment to pass through. One end of each group of drive lines is fixed to the support disk at the end of the arm segment through the base disk and the middle support disk, and the other end is wound around multiple winding wheels on the frame. A motor is connected to the winding wheels.

[0006] The beneficial effects of the above solution are as follows: This invention proposes a line-driven, continuous soft robot that uses an elastic tube as the main body of the robotic arm, undertaking the tasks of motion, support, and transport. By integrating the elastic tube liquid transfer mechanism and the continuous robot's elastic arm into a single design, this invention eliminates the need for a separate elastic support, resulting in a more compact robot structure and lower manufacturing costs. The robot is used for automatic cleaning of ships at docks. It controls the rotation of a motor to drive a winding wheel, which in turn moves the drive line. Each set of drive lines provides redundant drive to each arm segment of the continuous robot, enabling bending movements in each segment.

[0007] Furthermore, the drive line is made of a super-elastic alloy material.

[0008] The beneficial effects of the above-mentioned further solutions are: the structure becomes more compact and reliable, and can operate in narrow or irregular spaces.

[0009] In addition, the technical solution adopted in this invention is: a control method for a robot integrating an elastic tube liquid transfer mechanism and a continuous robot elastic arm, comprising the following steps:

[0010] S1: Based on the integration of the elastic tube liquid transfer mechanism and the continuous robot elastic arm, a pose error equation is constructed from the workspace to the arm segment space to correct the pose error during the robot's movement.

[0011] S2: Solve the virtual joint variables based on the pose error equation, establish the relationship between the arm segment space and the drive space, and make the robot move according to the desired pose by controlling the rotation of the motor;

[0012] S3: Based on the robot's movement, a structured light depth camera is used to acquire the environmental conditions of the area to be cleaned, and real-time online target detection is performed based on the YOLOv5 model. The robot position is adjusted in real time to complete the control of the robot that integrates the elastic tube liquid transfer mechanism and the continuous robot elastic arm.

[0013] The beneficial effects of the above solution are: analyzing the workspace-to-arm space and the arm-to-drive space separately allows for better correction of pose errors during robot motion solving, and better control of the robot to move according to the desired position and posture; real-time target detection using machine vision facilitates online correction and adjustment of the robotic arm's pose. This solution solves the problem that existing technologies cannot effectively correct pose errors during robot kinematics solving.

[0014] Furthermore, the pose error equation from the workspace to the arm segment space is constructed in S1, including the following steps:

[0015] S1-1: The arm segment is equivalent to a rod with uniform curvature, and the virtual joint angle vector q is defined as:

[0016]

[0017] in, and Let θ1, θ2, and θ3 be the bending plane angles of each arm segment, and T be the transpose of the matrix.

[0018] S1-2: Establish the kinematic pose error equation based on the virtual joint angle vector, using the pose tracking error e as the controlled variable and the joint angular velocity... As the control variable, the control objective is to find a control law that makes the pose tracking error e converge to zero within a specified time. The joint angular velocity is then solved using nonlinear control theory. The pose error equation from the workspace to the arm segment space is constructed. The kinematic pose error equation is as follows:

[0019]

[0020] Here, A(q) is a nonlinear function with respect to q, and the dot above the symbol indicates the derivative with respect to time.

[0021] The beneficial effects of the above-mentioned further solutions are: through the above technical solutions, the conversion from the workspace to the arm segment space is completed, and the pose error is corrected during the robot motion solution process.

[0022] Furthermore, the relationship between the arm segment space and the drive space established in S2 is as follows:

[0023] Change in rope length Δl of segment j in hole i i,j for:

[0024]

[0025] Where i = 1, 2, ..., 9, j = 1, 2, 3, and r is the distance between the through hole of the supporting disk and the center of the elastic tube. Let θ be the bending plane of arm segment j. j The bending angle of arm segment J;

[0026] The change in length Δl of the s-th rope relative to the case where the rope passes through the arm segment without bending. s for:

[0027] Δl1=Δl 1，1

[0028] Δl4=Δl 4，1

[0029] Δl7=Δl 7，1

[0030] Δl2=Δl 2，1 +Δl2，2

[0031] Δl5=Δl 5，1 +Δl 5，2

[0032] Δl8=Δl 8，1 +Δl 8，2

[0033] Δl3=Δl 3，1 +Δl 3，2 +Δl 3，3

[0034] Δl6=Δl 6，1 +Δl 6，2 +Δl 6，3

[0035] Δl9=Δl 9，1 +Δl 9，2 +Δl 9，3

[0036] s = 1, 2, ..., 9

[0037] For Δl i,j Differentiating the derivative, we obtain the linear velocity of the winding wheel as v. c :

[0038]

[0039] angular velocity ω of the winding wheel c for:

[0040] ω c =v c / r c

[0041] Where, r c The radius of the winding wheel is denoted as .

[0042] The beneficial effects of the above-mentioned further solutions are: by establishing the relationship between the arm segment space and the drive space through the above technical solutions, the transformation from the arm segment space to the drive space is completed, which is conducive to better control of the robot's movement.

[0043] Furthermore, S2 uses a microcontroller to control the motor rotation via a dual-loop PID method, including the following steps:

[0044] S2-1: Input the motor speed, obtain the motor speed using the motor encoder, calculate the speed deviation, and use the PID algorithm to calculate the output of the speed loop controller as the set value of the current loop;

[0045] S2-2: Calculate the deviation between the current setpoint and the actual current output of the speed loop, and use the PID algorithm to calculate the output of the current loop controller;

[0046] S2-3: The duty cycle is controlled by the output of the current loop, thereby controlling the motor speed.

[0047] The advantages of the above-mentioned further solutions are: by using a dual-loop PID controller to control the motor, it is easy to implement in engineering and has good stability.

[0048] Furthermore, S3 includes the following sub-steps:

[0049] S3-1: Collect the environmental condition dataset of the area to be cleaned, and perform image random transformation and mosaic preprocessing on the dataset;

[0050] S3-2: Train the YOLOv5 model using the preprocessed dataset and perform transfer learning using the pre-trained model on the VOC dataset;

[0051] S3-3: Based on the transfer learning results, target detection is performed using the trained YOLOv5 model, and two-dimensional coordinates are calculated using the anchor boxes drawn in the image;

[0052] S3-4: Based on two-dimensional coordinates, the depth map acquired by the structured light depth camera is registered with the RGB map to obtain the Cartesian space coordinates of the area to be cleaned;

[0053] S3-5: Transmits the Cartesian space coordinates and the position and spatial distance information of the area to be cleaned to the control board to adjust the robot position in real time.

[0054] The beneficial effects of the above-mentioned further solutions are as follows: Due to the complex working environment of continuous robots, traditional uniform-speed cleaning technology suffers from insufficient cleaning in this scenario, cannot adapt to environmental changes to adjust the flow rate, and cannot meet the needs of actual use. Therefore, this invention uses a structured light depth camera to acquire the environmental conditions of the area to be cleaned, and then uses a YOLOv5-based model for target detection, aligning the color image with the depth image to obtain the position and spatial distance information of the area to be cleaned, so as to adjust the position of the robotic arm in real time and improve the quality of operation. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of a robot structure that integrates an elastic tube liquid transfer mechanism with a continuous robot elastic arm.

[0056] The components include: 1. Frame; 2. Elastic tube; 3. Drive line; 4. Winding wheel; 5. Motor.

[0057] Figure 2 A flowchart illustrating the control method for a robot that integrates a flexible tube liquid transfer mechanism with a continuous robotic elastic arm.

[0058] Figure 3This is a schematic diagram of the bending deformation of the arm segment.

[0059] Figure 4 This is the flowchart for dual-loop PID control. Detailed Implementation

[0060] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0061] Example 1, as Figure 1 As shown, a robot integrating an elastic tube liquid transfer mechanism with a continuous robot elastic arm includes a frame 1, an elastic tube 2, a support disk, a base disk, and a drive line 3. The elastic tube 2 is fixed to the frame 1 and is divided into multiple arm segments. Each arm segment is sequentially provided with a base disk and multiple support disks, and the spacing between each support disk is equal. The base disk and support disks are fixed to the elastic tube 2 by clamps. The support disks have multiple through holes distributed on them. Every three through holes spaced at 120° intervals form a group for the drive line 3 of one arm segment to pass through. One end of each group of drive lines 3 is fixed to the support disk at the end of the arm segment through the base disk and the middle support disk. The other end is wound around multiple winding wheels 4 set on the frame 1. The winding wheels 4 are connected to motors 5.

[0062] Drive line 3 is made of a super-elastic alloy material.

[0063] In one embodiment of the invention, the elastic tube 2 has an outer diameter D and an inner diameter d, serving to form the overall shape of the robot, transport water, and provide the variable stiffness required for the robot's bending process. The total length of the tube can be adjusted according to different working environments. The tube is divided into three arm segments, each with 2 degrees of freedom. Each arm segment of the robot has 3 supporting disks, with equal spacing between the disks, and the radius of the supporting disks is r. a Nine through holes are evenly distributed on the pitch circle. Three through holes spaced 120° apart form a group for the passage of the superelastic alloy drive wires. The nine alloy drive wires are divided into three groups: the first group passes through the base disk and the intermediate support disk, then is fixed to the third support disk; the second group passes through the base disk and the intermediate support disk, then is fixed to the sixth support disk; and the third group passes through all the disks and is fixed to the end disk.

[0064] The portion between the base disk and the disk fixed to the first set of drive lines constitutes the first arm segment of the continuous robot, and the disk fixed to the drive lines is the end disk of this segment. The end disk of the first arm segment is also the base disk of the second arm segment, and the portion between the base disk of the second arm segment and the disk fixed to the second set of drive lines constitutes the second arm segment of the continuous robot; similarly, the end disk of the second arm segment is also the base disk of the third arm segment, and the portion between the base disk of the third arm segment and the end support disk constitutes the third arm segment of the continuous robot.

[0065] Example 2, as Figure 2 As shown, a control method for a robot integrating an elastic tube liquid transfer mechanism and a continuous robotic elastic arm includes the following steps:

[0066] S1: Based on the integration of the elastic tube liquid transfer mechanism and the continuous robot elastic arm, a pose error equation is constructed from the workspace to the arm segment space to correct the pose error during the robot's movement.

[0067] S2: Solve the virtual joint variables based on the pose error equation, establish the relationship between the arm segment space and the drive space, and make the robot move according to the desired pose by controlling the rotation of the motor;

[0068] S3: Based on the robot's movement, a structured light depth camera is used to acquire the environmental conditions of the area to be cleaned, and real-time online target detection is performed based on the YOLOv5 model. The robot position is adjusted in real time to complete the control of the robot that integrates the elastic tube liquid transfer mechanism and the continuous robot elastic arm.

[0069] The pose error equation from the workspace to the arm segment space is constructed in S1, including the following steps:

[0070] S1-1: The arm segment is equivalent to a rod with uniform curvature, such as... Figure 3 As shown, the virtual joint angle vector q is defined as:

[0071]

[0072] in, and Let θ1, θ2, and θ3 be the bending plane angles of each arm segment, and T be the transpose of the matrix.

[0073] Figure 3 China i-1 x i-1 y i-1 z i-1 o i x i y i z iThese are the fixed coordinate systems connected to the two support disks at the beginning and end of the i-th segment, with their origins at the centers of the disks at the two ends, respectively.

[0074] S1-2: Establish the kinematic pose error equation based on the virtual joint angle vector, using the pose tracking error e as the controlled variable and the joint angular velocity... As the control variable, the control objective is to find a control law that makes the pose tracking error e converge to zero within a specified time. The joint angular velocity is then solved using nonlinear control theory. The pose error equation from the workspace to the arm segment space is constructed. The kinematic pose error equation is as follows:

[0075]

[0076] Here, A(q) is a nonlinear function with respect to q, and the dot above the symbol indicates the derivative with respect to time.

[0077] The relationship between the arm segment space and the drive space in S2 is established as follows:

[0078] Change in rope length Δl of segment j in hole i i,j for:

[0079]

[0080] Where i = 1, 2, ..., 9, j = 1, 2, 3, and r is the distance between the through hole of the supporting disk and the center of the elastic tube. Let θ be the bending plane of arm segment j. j The bending angle of arm segment J;

[0081] The change in length Δl of the s-th rope relative to the case where the rope passes through the arm segment without bending. s for:

[0082] Δl1=Δl 1，1

[0083] Δl4=Δl 4，1

[0084] Δl7=Δl 7，1

[0085] Δl2=Δl 2，1 +Δl 2，2

[0086] Δl5=Δl 5，1 +Δl 5，2

[0087] Δl8=Δl 8，1 +Δl 8，2

[0088] Δl3=Δl 3，1 +Δl3，2 +Δl 3，3

[0089] Δl6=Δl 6，1 +Δl 6，2 +Δl 6，3

[0090] Δl9=Δl 9，1 +Δl 9，2 +Δl 9，3

[0091] s = 1, 2, ..., 9

[0092] For Δl i,j Differentiating the derivative, we obtain the linear velocity of the winding wheel as v. c :

[0093]

[0094] angular velocity ω of the winding wheel c for:

[0095] ω c =v c / r c

[0096] Where, r c The radius of the winding wheel is denoted as .

[0097] S2 uses a microcontroller to control the motor rotation via a dual-loop PID method, such as... Figure 4 As shown, it includes the following steps:

[0098] S2-1: Input the motor speed, obtain the motor speed using the motor encoder, calculate the speed deviation, and use the PID algorithm to calculate the output of the speed loop controller as the set value of the current loop;

[0099] S2-2: Calculate the deviation between the current setpoint and the actual current output of the speed loop, and use the PID algorithm to calculate the output of the current loop controller;

[0100] S2-3: The duty cycle is controlled by the output of the current loop, thereby controlling the motor speed.

[0101] S3 includes the following steps:

[0102] S3-1: Collect the environmental condition dataset of the area to be cleaned, and perform image random transformation and mosaic preprocessing on the dataset to prevent overfitting;

[0103] S3-2: Train the YOLOv5 model using the preprocessed dataset and perform transfer learning using the pre-trained model on the VOC dataset;

[0104] S3-3: Based on the transfer learning results, target detection is performed using the trained YOLOv5 model, and two-dimensional coordinates are calculated using the anchor boxes drawn in the image;

[0105] S3-4: Based on two-dimensional coordinates, the depth map acquired by the structured light depth camera is registered with the RGB map to obtain the Cartesian space coordinates of the area to be cleaned;

[0106] S3-5: Transmits the Cartesian space coordinates and the position and spatial distance information of the area to be cleaned to the control board to adjust the robot position in real time.

[0107] The elastic tube of this invention includes a liquid transfer mechanism, integrating the elastic tube liquid transfer mechanism and the continuous robotic elastic arm into a single design. This eliminates the need for a separate elastic support, resulting in a more compact robot structure and lower manufacturing costs. Using the elastic tube as the main body of the robotic arm and driving it via a linear drive system, the structure is simple and reliable. Compared to traditional robotic arms, it better handles tasks in small or complex workspaces. This invention constructs a pose error equation from the workspace to the arm segment space, which can better correct pose errors during robot motion solving, improving the robot's tracking accuracy. This invention establishes a relationship between the arm segment space and the drive space, allowing the robotic arm to move to the desired position and posture simply by controlling the motor rotation. Simultaneously, dual-loop PID control of the motor facilitates engineering implementation and provides good stability. This invention establishes a target detection mechanism that continuously feeds back target area position information, improving control accuracy. Furthermore, it can accurately acquire target area information even in blind spots, enhancing operational quality.

[0108] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of the invention.

Claims

1. A control method for a robot integrating an elastic tube liquid transfer mechanism and a continuous robot elastic arm, wherein the robot integrating the elastic tube liquid transfer mechanism and the continuous robot elastic arm includes a frame (1), an elastic tube (2), a support disk, a base disk, and a drive line (3). The elastic tube (2) is fixed on the frame (1). The elastic tube (2) is divided into multiple arm segments. Each arm segment is sequentially provided with a base disk and multiple support disks, and the spacing between each support disk is equal. All the support discs are fixed to the elastic tube (2) by clamps. The support discs have multiple through holes. Every three through holes spaced at 120° are grouped together for the passage of the drive line (3) of one arm segment. One end of each group of drive lines (3) is fixed to the support disc at the end of the arm segment by the base disc and the middle support disc. The other end is wound around multiple winding wheels (4) on the frame (1). The winding wheels (4) are connected to the motor (5). The drive line (3) is made of super-elastic alloy material. Its features are, The control method includes the following steps: S1: Based on the integration of the elastic tube liquid transfer mechanism and the continuous robot elastic arm, a pose error equation is constructed from the workspace to the arm segment space to correct the pose error during the robot's movement. The pose error equation from the workspace to the arm segment space is constructed in S1, including the following steps: S1-1: The arm segment is equivalent to a rod with uniform curvature, and a virtual joint angle vector is defined. for: in, , and The bending plane angle of each arm segment, , and The bending angle of each arm segment, This is the transpose of the matrix; S1-2: Establish the kinematic pose error equation based on the virtual joint angle vector, and incorporate pose tracking error. As a controlled variable, joint angular velocity As the control variable, the control objective is to find a control law that minimizes the pose tracking error. The joint angular velocity converges to zero within a specified time, and is then calculated using nonlinear control theory. The pose error equation from the workspace to the arm segment space is constructed, and the kinematic pose error equation is as follows: in, For about The nonlinear function is indicated by the dot above the sign, which represents the derivative with respect to time. S2: Solve the virtual joint variables based on the pose error equation, establish the relationship between the arm segment space and the drive space, and make the robot move according to the desired pose by controlling the rotation of the motor; The relationship between the arm segment space and the drive space established in S2 is as follows: No. Kong Di Change in arm length for: in, , , To support the distance between the through hole of the disc and the center of the elastic tube, for Arm segment bending plane, for Arm segment bending angle; No. The change in rope length relative to the straight section of the arm. for: right Differentiating the derivative, we obtain the linear velocity of the winding wheel as follows: : angular velocity of the winding wheel for: in, The radius of the winding wheel; S3: Based on the robot's movement, a structured light depth camera is used to acquire the environmental conditions of the area to be cleaned, and real-time online target detection is performed based on the YOLOv5 model. The robot position is adjusted in real time to complete the control of the robot that integrates the elastic tube liquid transfer mechanism and the continuous robot elastic arm.

2. The control method for a robot integrating the elastic tube liquid transfer mechanism and the continuous robotic elastic arm according to claim 1, characterized in that, In step S2, a microcontroller is used to control the motor rotation via a dual-loop PID method, including the following steps: S2-1: Input the motor speed, obtain the motor speed using the motor encoder, calculate the speed deviation, and use the PID algorithm to calculate the output of the speed loop controller as the set value of the current loop; S2-2: Calculate the deviation between the current setpoint and the actual current output of the speed loop, and use the PID algorithm to calculate the output of the current loop controller; S2-3: The duty cycle is controlled by the output of the current loop, thereby controlling the motor speed.

3. The control method for a robot integrating the elastic tube liquid transfer mechanism and the continuous robotic elastic arm according to claim 1, characterized in that, S3 includes the following sub-steps: S3-1: Collect the environmental condition dataset of the area to be cleaned, and perform image random transformation and mosaic preprocessing on the dataset; S3-2: Train the YOLOv5 model using the preprocessed dataset and perform transfer learning using the pre-trained model on the VOC dataset; S3-3: Based on the transfer learning results, target detection is performed using the trained YOLOv5 model, and two-dimensional coordinates are calculated using the anchor boxes drawn in the image; S3-4: Based on two-dimensional coordinates, the depth map acquired by the structured light depth camera is registered with the RGB map to obtain the Cartesian space coordinates of the area to be cleaned; S3-5: Transmits the Cartesian space coordinates and the position and spatial distance information of the area to be cleaned to the control board to adjust the robot position in real time.

Images