Humanoid robot welding method and device and humanoid robot

Abstract

The invention relates to the technical field of robot control, and provides a humanoid robot welding method and device and a humanoid robot. The method comprises the steps that in the walking welding process of the humanoid robot, walking parameters of the humanoid robot at the next moment are determined based on the gravity center variation of the humanoid robot at the current moment; in the walking welding process of the humanoid robot, the pose of the welding mechanical arm at the next moment is determined based on the welding track deviation of the welding mechanical arm at the current moment, the walking parameters of the welding mechanical arm at the next moment, the pose of the welding mechanical arm at the current moment and a preset welding track; and based on the pose at the next moment, the welding mechanical arm is controlled to conduct welding. According to the method, the welding track can be accurately controlled, it is guaranteed that the welding arc is always located on the center line of the welding seam, the welding quality is improved, and then the welding precision, continuity and welding quality of the humanoid robot on the long welding seam are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control, and particularly to a humanoid robot welding method, device and humanoid robot. Background Art

[0002] Humanoid robot welding refers to the process of applying a humanoid robot to welding operations and using its flexibility and autonomy to complete welding tasks in a complex environment.

[0003] Although humanoid robots can imitate human movements and have certain autonomous navigation and path planning capabilities, during the welding process of current humanoid robots, due to the complexity of their own movements and the interference of the external environment, it is easy to cause the actual welding trajectory to deviate from the target weld seam, thereby triggering quality problems such as discontinuous welding joints, poor welding formation, or uneven welding strength. Summary of the Invention

[0004] The present invention provides a humanoid robot welding method, device and humanoid robot to solve the defects existing in the prior art.

[0005] The present invention provides a humanoid robot welding method, including the following steps: During the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment, where the change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; During the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding manipulator at the current moment, the walking parameters at the next moment, the pose of the welding manipulator at the current moment, and a preset welding trajectory, determine the pose of the welding manipulator at the next moment, where the welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding manipulator at the current moment and the center line of the target weld seam; Based on the pose at the next moment, control the welding manipulator to weld.

[0006] According to the humanoid robot welding method provided by the present invention, the step of determining the walking parameters of the humanoid robot at the next moment based on the change amount of the center of gravity of the humanoid robot at the current moment includes: Obtain the motion state of the humanoid robot at the current moment; Input the motion state at the current moment and the change amount of the center of gravity at the current moment into a reinforcement learning model to obtain the walking parameters at the next moment output by the reinforcement learning model; The reinforcement learning model is trained with the goal of minimizing the reward value, and the reward value is determined based on the change amount of the center of gravity of the humanoid robot.

[0007] According to a humanoid robot welding method provided by the present invention, determining the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and a preset welding trajectory includes: Based on the preset welding trajectory, a dynamic model during the walking process of the humanoid robot is constructed. The dynamic model is used to describe the motion relationship and force conditions among various components during the walking process of the humanoid robot, as well as the kinematic constraints of the welding robotic arm; Based on the dynamic model, the welding trajectory deviation at the current moment, the walking parameters at the next moment, and the pose of the welding robotic arm at the current moment, determine the optimal adjustment path and pose compensation amount of the welding robotic arm; Based on the optimal adjustment path and the pose compensation amount, determine the pose of the welding robotic arm at the next moment.

[0008] According to a humanoid robot welding method provided by the present invention, the steps for obtaining the motion state and pose at the current moment are as follows: Collect the initial motion state and the initial pose at the current moment through sensors; Filter the initial motion state at the current moment to obtain the motion state at the current moment; Filter the initial pose at the current moment to obtain the pose at the current moment.

[0009] According to a humanoid robot welding method provided by the present invention, it further includes: When the change amount of the center of gravity at the current moment is greater than a first threshold, and / or the welding trajectory deviation at the current moment is greater than a second threshold, a warning prompt is given.

[0010] The present invention also provides a humanoid robot welding device, including the following modules: A first determination unit, configured to determine the walking parameters of the humanoid robot at the next moment based on the change amount of the center of gravity of the humanoid robot at the current moment during the walking and welding process of the humanoid robot. The change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; A second determination unit, configured to determine the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and a preset welding trajectory during the walking and welding process of the humanoid robot. The welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robotic arm at the current moment and the center line of the target weld; A welding unit for controlling the welding of the welding robot arm based on the pose at the next moment.

[0011] The present invention also provides a humanoid robot, comprising: a robot body, a welding robot arm installed on the robot body, and the humanoid robot welding device as described above.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the method for welding a humanoid robot as described in any one of the above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for welding a humanoid robot as described in any one of the above is implemented.

[0014] The present invention also provides a computer program product, comprising a computer program. When the computer program is executed by a processor, the method for welding a humanoid robot as described in any one of the above is implemented.

[0015] The method, device and humanoid robot for welding a humanoid robot provided by the present invention realize the balance control of the walking process of the humanoid robot by real-time monitoring the change amount of the center of gravity of the humanoid robot and dynamically adjusting the walking parameters, ensuring that the humanoid robot remains stable during the walking process and providing a reliable platform for the welding operation. In addition, the present invention also precisely controls the pose of the welding robot arm at the next moment by comprehensively considering the welding trajectory deviation, walking parameters, the pose of the welding robot arm at the current moment and the preset welding trajectory, realizing the precise control of the welding trajectory, ensuring that the welding arc is always located on the weld center line, improving the welding quality, and further ensuring the accuracy, continuity and welding quality of the humanoid robot for welding long welds. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 It is a schematic flowchart of the method for welding a humanoid robot provided by the present invention.

[0018] Figure 2 It is a schematic flowchart of the center of gravity control of the humanoid robot provided by the present invention.

[0019] Figure 3 It is a schematic flowchart of the pose adjustment of the welding robot arm provided by the present invention.

[0020] Figure 4 It is a schematic structural diagram of the humanoid robot welding device provided by the present invention.

[0021] Figure 5 It is a schematic structural diagram of the electronic device provided by the present invention. Detailed implementation manners

[0022] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0023] Currently, in the long-weld welding operation, it is difficult for a humanoid robot to meet the requirements of walking and welding at the same time. The main problems are as follows: (1) Dynamic balance problem: During the walking process of the robot, the position of the center of gravity continuously changes, which easily causes phenomena such as vibration, inclination or instability, and further leads to the deviation of the weld seam during the welding process, affecting the welding quality. (2) It is difficult to accurately maintain the welding trajectory: When the robot walks, the welding manipulator needs to synchronously maintain a relatively stable posture with respect to the weld seam. However, there are deficiencies in accurately maintaining the welding trajectory during the walking process at present, which easily causes problems such as discontinuous welding joints or uneven welding quality.

[0024] In view of this, the present invention provides a humanoid robot welding method to solve the problems that it is difficult to maintain dynamic balance and difficult to accurately maintain the welding trajectory during the walking and welding process of the humanoid robot.

[0025] Among them, Figure 1 It is a schematic flow diagram of the humanoid robot welding method provided by the present invention. As Figure 1 shown, the method includes step 110, step 120 and step 130.

[0026] Step 110: During the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment. The change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment.

[0027] Here, the change in the center of gravity at the current moment refers to the vector difference between the center of gravity position of the humanoid robot at the current moment and that at the previous moment, which is used to measure the balance state and movement trend of the humanoid robot. The larger the change in the center of gravity at the current moment, the more unstable the current balance state of the humanoid robot, and there is a risk of tipping over or significant vibration. At this time, the walking parameters need to be adjusted to quickly restore the balance state and reduce the center of gravity offset. Among them, the change in the center of gravity at the current moment can be directly measured and obtained by an inertial measurement unit (IMU) installed on the robot, or can be obtained by calculating the kinematic data of each joint of the robot. For example, based on the kinematic data, the centroid positions of each component of the humanoid robot can be calculated using the forward kinematic formula, and then the center of gravity position at the current moment can be obtained by weighted averaging, and the difference from the previous moment can be calculated.

[0028] The walking parameters for the next moment refer to the set of instructions for controlling the movement state of the humanoid robot in the next moment. This set of instructions can include parameters such as step length, walking speed, gait cycle, support phase duration, swing phase trajectory, hip height, and ankle angle. The walking parameters are determined based on the change in the center of gravity at the current moment, that is, they are set to offset or reduce the current trend of change in the center of gravity and maintain the balance state of the robot.

[0029] As an alternative embodiment, the step length and walking speed for the next step can be adjusted based on the magnitude of the change in the center of gravity at the current moment. When the change in the center of gravity is large, the step length and walking speed can be reduced, and the support time of the supporting leg can be increased to improve stability; when the change in the center of gravity is small, the step length and walking speed can be increased to improve walking efficiency.

[0030] Since the embodiments of the present invention consider the change in the center of gravity when determining the walking parameters, it can ensure that the humanoid robot can timely adjust its movement posture during walking, maintain balance, avoid tipping over, and ensure the dynamic balance of the humanoid robot during walking and welding.

[0031] Step 120: During the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding robot arm at the current moment, the walking parameters for the next moment, the pose of the welding robot arm at the current moment, and the preset welding trajectory, determine the pose of the welding robot arm for the next moment. The welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robot arm at the current moment and the center line of the target weld.

[0032] Specifically, the actual welding trajectory at the current moment refers to the path that the welding arc or the end of the welding torch on the welding robot arm actually moves on the surface of the welded workpiece during the welding process. The center line of the target weld is the center line of the weld preset in advance, usually an ideal welding path determined in advance according to the geometric shape of the welded workpiece and the welding requirements. The larger the deviation between the two, the worse the welding quality, and it is easy to appear welding defects, and even lead to welding failure.

[0033] A welding robot arm refers to a robot arm installed on a humanoid robot for performing welding tasks. The current position refers to the position and posture of the end effector of the welding robot arm (usually a welding gun) in three-dimensional space, including position coordinates (X, Y, Z) and posture angles (such as Euler angles or quaternions). The preset welding trajectory refers to a pre-planned welding path based on the geometric shape of the welding workpiece and the welding process requirements, usually in the form of a series of ordered three-dimensional space points or curves.

[0034] Considering that the humanoid robot has changes in center of gravity and motion disturbances during walking, the motion accuracy of the welding robot arm is affected by the movement of the humanoid robot. There may be interference from the external environment during the welding process, such as wind force, temperature changes, etc., and the welding trajectory deviation at the current moment reflects the error of the current welding process, which is an important basis for adjusting the posture of the robot arm. The walking parameters at the next moment affect the overall motion trend of the humanoid robot and have a constraining effect on the posture adjustment of the welding robot arm. The posture of the welding robot arm at the current moment is the basis for calculating the posture at the next moment, and provides the current spatial state information of the robot arm. The preset welding trajectory gives an ideal welding path and is the target reference for posture adjustment. Therefore, the embodiment of the present invention determines the posture of the welding robot arm at the next moment based on the welding trajectory deviation at the current moment, the walking parameters at the next moment, the posture of the welding robot arm at the current moment, and the preset welding trajectory, so as to ensure that the actual welding trajectory of the welding robot arm coincides with the target weld as much as possible, reduce the welding deviation, and thus ensure efficient and stable long weld mobile welding operations in complex working environments, thereby improving the quality of welds.

[0035] Step 130: Based on the posture at the next moment, control the welding robot arm to weld.

[0036] Specifically, the posture at the next moment refers to the position and posture that the end effector (welding gun) of the welding robot arm should reach in the next control cycle, which is used to guide the movement of the welding robot arm. Since the posture at the next moment is determined based on the welding trajectory deviation at the current moment, the walking parameters at the next moment, the posture of the welding robot arm at the current moment, and the preset welding trajectory, it can ensure that the welding robot arm eliminates the welding trajectory deviation as much as possible during the movement, fits the preset welding trajectory, and coordinates with the walking movement of the humanoid robot, so that when the welding robot arm is controlled to perform welding based on the posture at the next moment, it can achieve accurate welding trajectory control, ensure welding quality, and improve welding efficiency.

[0037] As an optional embodiment, a PID controller or other motion control algorithm can be used to calculate the control instructions of each joint of the welding robot arm according to the posture at the next moment, drive the motor to move, and make the welding gun accurately reach the target position and posture.

[0038] The humanoid robot welding method provided by the embodiment of the present invention realizes the balance control of the walking process of the humanoid robot by real-time monitoring the change amount of the center of gravity of the humanoid robot and dynamically adjusting the walking parameters, ensuring that the humanoid robot remains stable during the walking process and providing a reliable platform for the welding operation. In addition, the embodiment of the present invention also accurately controls the pose of the welding robotic arm at the next moment by comprehensively considering the welding trajectory deviation, walking parameters, the pose of the welding robotic arm at the current moment, and the preset welding trajectory, realizes the accurate control of the welding trajectory, ensures that the welding arc is always located on the weld center line, improves the welding quality, and further ensures the accuracy, continuity, and welding quality of the humanoid robot for long weld welding.

[0039] Based on the above embodiment, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment, including: Obtain the motion state of the humanoid robot at the current moment; Input the motion state at the current moment and the change amount of the center of gravity at the current moment into the reinforcement learning model, and obtain the walking parameters at the next moment output by the reinforcement learning model; The reinforcement learning model is trained with the goal of minimizing the reward value, and the reward value is determined based on the change amount of the center of gravity of the humanoid robot.

[0040] Specifically, the motion state at the current moment refers to the information set describing the current motion posture and speed of the humanoid robot, which may include parameters such as the angles, angular velocities, angular accelerations of each joint, the posture of the torso, linear velocity, angular velocity, center of gravity position, and speed. Among them, the motion state at the current moment can be directly measured and obtained by sensors (such as encoders, gyroscopes, accelerometers, etc.) installed on the humanoid robot, or can be obtained by calculating the kinematic data of each joint of the robot. The embodiment of the present invention does not make specific limitations on this.

[0041] In addition, the reinforcement learning model refers to an intelligent decision-making model based on trial-and-error learning. By interacting with the environment, it continuously learns and optimizes the strategy to achieve the set goal. It is trained with the goal of minimizing the reward value, and the reward value is determined based on the change amount of the center of gravity of the humanoid robot. That is to say, the goal of the reinforcement learning model is to learn a walking strategy that enables the humanoid robot to maintain balance as much as possible during the walking process, reduce the center of gravity offset, and thus obtain a higher reward.

[0042] Among them, the reinforcement learning model can be trained and optimized using algorithms such as the Proximal Policy Optimization algorithm (PPO), the Deep Deterministic Policy Gradient algorithm (DDPG), etc., so that it can output appropriate walking parameters at the next moment according to the motion state and the change amount of the center of gravity at the current moment.

[0043] For example, the state space and action space can be designed. The state space includes parameters such as the current posture, center of gravity position, speed, and acceleration of the humanoid robot, and the action space is the control instructions for each drive motor. The center-of-gravity regulation strategy during the robot's walking process is trained online using deep reinforcement learning algorithms (such as DDPG, PPO, etc.) to enable the humanoid robot to achieve stable movement of the center of gravity in a complex dynamic environment. During the training process, a reward function is introduced, and the reward function comprehensively considers multiple indicators such as the balance of the robot (for example, the degree of deviation of the center of gravity from the center position of the weld), energy consumption, motion smoothness, and welding trajectory stability. Among them, the design of the reward function needs to be adjusted according to the actual application scenario to enable the robot to achieve adaptive center-of-gravity balance control during walking and achieve the best balance control effect.

[0044] Figure 2 is a schematic diagram of the center-of-gravity control process of the humanoid robot provided by the present invention. As Figure 2 shown, first, the current welding posture of the humanoid robot, including information such as welding pose, center of gravity, and speed, is collected in real time through sensors, and at the same time, welding process parameters (voltage, current, speed, swing amplitude) are collected as input data. Then, the current state (S) is input into the policy (Actor) network, and the policy network outputs the corresponding action (A) according to the current state, that is, adjusts the walking parameters of the humanoid robot to optimize the welding posture and maintain balance. Among them, the current state (S) includes the current welding posture of the humanoid robot and the welding process parameters.

[0045] At the same time, the target state (S’) is input into the policy (Actor) target network to obtain the target action (A’). In addition, the current state (S) is processed by noise (Noise) and then enters the value (Critic) target network for the next network training. Meanwhile, the value (Critic) network (including value network 1 and value network 2) receives the current state (S) and action (A). Value network 1 calculates the current value 1 (Q1) based on the current state (S) and action (A), and value network 2 calculates the current value 2 (Q2) based on the current state (S) and action (A). Q1 and Q2 are used to evaluate the quality of the current policy. Then, the error 1 (such as the temporal difference error TD-error) is calculated to guide the parameter update of value network 1, and the error 2 (such as the temporal difference error TD-error) is calculated to guide the parameter update of value network 2 to improve the accuracy of value estimation.

[0046] Then, the value target network 1 and the value target network 2 calculate the target value (Q') based on the target state (S') and the target action (A'), compare it with the current value Q (Q1 and Q2), and further optimize the network. This process forms a closed-loop Actor-Critic reinforcement learning framework. Through continuous iterative training, the policy network can output more optimal walking parameters to ensure that the humanoid robot can effectively adjust the center-of-gravity change during the walking and welding process and maintain dynamic balance.

[0047] Finally, the walking parameters output by the trained policy network are used to actually control the humanoid robot, realizing real-time adjustment of the robot's movement, ensuring the stability of the welding posture and the improvement of the welding quality. The entire process takes the center-of-gravity state feedback as an important basis, continuously optimizes the robot's walking strategy, and realizes efficient and stable welding operations.

[0048] Based on any one of the above embodiments, determining the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and the preset welding trajectory includes: Based on the preset welding trajectory, construct a dynamic model during the walking process of the humanoid robot. The dynamic model is used to describe the motion relationship and force conditions among the components during the walking process of the humanoid robot, as well as the kinematic constraints of the welding robotic arm; Based on the dynamic model, the welding trajectory deviation at the current moment, the walking parameters at the next moment, and the pose of the welding robotic arm at the current moment, determine the optimal adjustment path and pose compensation amount of the welding robotic arm; Based on the optimal adjustment path and the pose compensation amount, determine the pose of the welding robotic arm at the next moment.

[0049] Specifically, the dynamic model refers to a mathematical model that describes the rigid-body dynamic characteristics, joint moments of inertia, mass distribution, mechanical constraints, and kinematic chain relationship of the humanoid robot and its welding robotic arm, and is used to simulate the dynamic response of the system under the force and motion states. As an alternative embodiment, the desired trajectories of the robot's joints can be discretized into a series of reference positions and velocity data at time steps according to the preset welding trajectory. Combining the robot configuration and physical parameters, the Lagrangian method or the Newton-Euler method can be used to establish the dynamic equation to form a complete motion constraint and force relationship model.

[0050] The optimal adjustment path refers to a continuous and smooth welding trajectory adjustment plan planned by the welding robotic arm to minimize the welding trajectory deviation and meet the kinematic and dynamic constraints under the influence of the current state and external disturbances. The pose compensation amount refers to the displacement and attitude rotation compensation values required by the welding robotic arm relative to the preset trajectory at the next moment to correct the trajectory deviation.

[0051] Considering that the center of gravity of a humanoid robot changes and dynamic disturbances occur during walking, the accuracy of the welding manipulator is easily affected. The dynamic model accurately describes the motion and force characteristics of each component of the robot. The welding trajectory deviation at the current moment reflects the error between the actual trajectory and the target trajectory. The walking parameters at the next moment reflect the overall motion trend of the robot. The pose of the welding manipulator at the current moment provides the current spatial state information. Therefore, based on the dynamic model, the welding trajectory deviation at the current moment, the walking parameters at the next moment, and the pose of the welding manipulator at the current moment, the embodiments of the present invention determine the optimal adjustment path and pose compensation amount of the welding manipulator, so as to effectively compensate for the welding trajectory error, ensure the continuity and accuracy of the motion of the welding manipulator, and improve the overall quality and stability of the weld.

[0052] After determining the optimal adjustment path and pose compensation amount, the compensated target pose can be converted into control commands for each joint of the manipulator through a motion control algorithm (such as model predictive control MPC or inverse kinematics solution) to determine the pose of the welding manipulator at the next moment and achieve accurate welding trajectory tracking.

[0053] Figure 3 It is a schematic flow chart of the pose adjustment of the welding manipulator provided by the present invention. As Figure 3 shown, first, obtain the motion target of the humanoid robot and the current pose of the humanoid robot through the state machine (including the position, attitude parameters of the end effector and the angles of each joint of the robot), and input the motion target, the current pose, and the pose of the welding manipulator obtained by forward kinematics into the trajectory planning module to obtain the expected trajectory planning parameters. At the same time, input the motion target and the current pose into the model predictive control (IS-MPC) module to obtain the optimized trajectory control quantity. Among them, forward kinematics refers to the pose of the welding manipulator calculated according to the joint parameters of the humanoid robot.

[0054] Input the expected trajectory parameters output by the trajectory planning module and the trajectory control quantity output by the model predictive control module into the trajectory processing module to obtain the final trajectory control command that is smooth and conforms to the constraints. Then, input the control command output by the trajectory processing module into the kinematics controller to obtain the target angles of each joint of the robot, and control the motion of each joint of the humanoid robot based on the target angles to achieve the accurate pose adjustment of the welding manipulator.

[0055] In addition, the state of the humanoid robot is detected in real time. The pose of the welding manipulator is obtained in real time through the information acquisition module. After being processed by the Kalman filter, the current pose of the humanoid robot is calculated by using inverse kinematics, and the pose of the humanoid robot is adjusted based on the difference between the current pose and the desired pose of the humanoid robot to ensure stable welding of the robot in a dynamic environment.

[0056] It can be seen that through trajectory planning and dynamic trajectory adjustment based on model predictive control, combined with real-time attitude feedback control of sensor acquisition and inverse kinematics, the embodiments of the present invention achieve high-precision active adjustment and trajectory maintenance of the pose of the welding manipulator of the humanoid robot, improving the stability of the welding process and the weld quality.

[0057] Based on any of the above embodiments, the steps for obtaining the motion state and pose at the current moment are as follows: Collect the initial motion state and initial pose at the current moment through sensors; Filter the initial motion state at the current moment to obtain the motion state at the current moment; Filter the initial pose at the current moment to obtain the pose at the current moment.

[0058] Specifically, the initial motion state and initial pose at the current moment collected through sensors may have large signal noise, vibration interference, and measurement errors due to the walking of the humanoid robot, resulting in unstable or abnormally fluctuating data. In this regard, the embodiments of the present invention filter the initial motion state and initial pose at the current moment respectively to obtain the motion state and pose at the current moment, so as to ensure that the acquired data is smoother and more accurate, and improve the stability and reliability of subsequent control algorithms.

[0059] Based on any of the above embodiments, the method further includes: Give a warning prompt when the center-of-gravity change amount at the current moment is greater than the first threshold and / or the welding trajectory deviation at the current moment is greater than the second threshold.

[0060] Specifically, if the center-of-gravity change amount at the current moment is greater than the first threshold, it indicates that the humanoid robot may be at risk of imbalance or tipping. In order to ensure the safety of the robot's walking and the continuity of welding, a warning prompt can be given.

[0061] If the welding trajectory deviation at the current moment is greater than the second threshold, it indicates that the actual welding trajectory of the welding manipulator deviates greatly from the target weld, which may lead to a decline in weld quality. In order to ensure welding quality and process stability, a warning prompt can be given.

[0062] It can be seen that through real-time monitoring of the robot's center-of-gravity change amount and welding trajectory deviation, and combining threshold judgment to implement an early warning mechanism, the embodiments of the present invention achieve dual monitoring and risk prevention of the robot's walking balance and welding pose, ensuring the safe and stable operation of the humanoid robot during welding operations and reliable guarantee of welding quality.

[0063] The humanoid robot welding device provided by the present invention will be described below. The humanoid robot welding device described below can be referred to in correspondence with the humanoid robot welding method described above.

[0064] Based on any of the above embodiments, Figure 4 is a schematic structural diagram of the humanoid robot welding device provided by the present invention. As Figure 4 shown, the device includes: A first determination unit 410, configured to determine the walking parameters of the humanoid robot at the next moment based on the change amount of the center of gravity of the humanoid robot at the current moment during the walking and welding process of the humanoid robot. The change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; A second determination unit 420, configured to determine the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and a preset welding trajectory during the walking and welding process of the humanoid robot. The welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robotic arm at the current moment and the center line of the target weld; A welding unit 430, configured to control the welding of the welding robotic arm based on the pose at the next moment.

[0065] Based on any of the above embodiments, determining the walking parameters of the humanoid robot at the next moment based on the change amount of the center of gravity of the humanoid robot at the current moment includes: Obtaining the motion state of the humanoid robot at the current moment; Inputting the motion state at the current moment and the change amount of the center of gravity at the current moment into a reinforcement learning model to obtain the walking parameters at the next moment output by the reinforcement learning model; The reinforcement learning model is trained with the goal of minimizing the reward value, and the reward value is determined based on the change amount of the center of gravity of the humanoid robot.

[0066] Based on any of the above embodiments, determining the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and a preset welding trajectory includes: Constructing a dynamic model during the walking process of the humanoid robot based on the preset welding trajectory. The dynamic model is used to describe the motion relationship and force condition between various components during the walking process of the humanoid robot, as well as the kinematic constraints of the welding robotic arm; Determining the optimal adjustment path and pose compensation amount of the welding robotic arm based on the dynamic model, the welding trajectory deviation at the current moment, the walking parameters at the next moment, and the pose of the welding robotic arm at the current moment; Determining the pose of the welding robotic arm at the next moment based on the optimal adjustment path and pose compensation amount.

[0067] Based on any of the above embodiments, the steps for obtaining the motion state and pose at the current moment are as follows: Collect the initial motion state and initial pose at the current moment through sensors; Filter the initial motion state at the current moment to obtain the motion state at the current moment; Filter the initial pose at the current moment to obtain the pose at the current moment.

[0068] Based on any of the above embodiments, it further includes: In the case where the change in the center of gravity at the current moment is greater than the first threshold and / or the deviation of the welding trajectory at the current moment is greater than the second threshold, a warning prompt is given.

[0069] Based on any of the above embodiments, the present invention further provides a humanoid robot, including: A robot body, a welding robotic arm installed on the robot body, and a humanoid robot welding device as described in any of the above embodiments.

[0070] Among them, the humanoid robot can be designed with a multi-legged or bipedal walking structure and has an adaptive gait adjustment function. Inside the robot chassis, high-performance servo motors, a drive system, and multi-degree-of-freedom sensors (including gyroscopes, accelerometers, and laser range sensors, etc.) are installed to monitor the posture changes and center-of-gravity offsets during the robot's walking in real time. The motion control system of the robot chassis realizes high-speed data transmission and control instruction interaction with the central processing unit, so as to adjust the angles of each joint in real time during walking to keep the robot stable.

[0071] The welding robotic arm is installed on the upper part of the robot and has the ability to move with multiple joints and multiple degrees of freedom. A welding torch and welding tooling are installed at the front end of the welding robotic arm, and various welding operations can be completed. In order to achieve high-precision tracking of the welding trajectory, the welding robotic arm can adopt a high-precision servo drive system and a feedback sensor to fine-tune the pose in real time under closed-loop control. At the same time, a high-resolution vision sensor and a laser sensor are built into the welding robotic arm to detect the real-time deviation information of the welding trajectory.

[0072] Based on any of the above embodiments, the present invention further provides a humanoid robot welding system, which includes: a humanoid robot, a data acquisition and processing module, a communication interface, and a fault self-diagnosis module.

[0073] Among them, the data acquisition and processing module includes a multi-channel sensor signal acquisition circuit, an A / D converter, and a signal filtering and processing unit. This module integrates the sensor data collected by each component of the robot chassis and the welding manipulator in real time, eliminates noise and interference through Kalman filtering and multi-sensor data fusion algorithms, and transmits the processed data to the humanoid robot welding device in the humanoid robot, providing an accurate basis for subsequent control decisions.

[0074] The humanoid robot welding device in the humanoid robot is used to train the optimal center of gravity control strategy of the humanoid robot in different motion states by using a deep reinforcement learning algorithm (such as DDPG). First, the system conducts offline simulation on the humanoid robot, constructs a state space and an action space model, and comprehensively evaluates the balance state, energy consumption, and motion smoothness of the robot through a reward function. During online operation, the humanoid robot inputs real-time state data into the reinforcement learning module to obtain corresponding control actions, and adjusts the robot's motion through a servo driver. Second, it is used for the welding manipulator based on the model predictive control (MPC) method. In each sampling period, this module calculates the optimal pose adjustment amount of the manipulator in the next period of time according to the welding trajectory deviation, the robot walking speed, the current position of the manipulator, and the preset welding trajectory, and feeds back the compensation instruction to the manipulator actuator to continuously correct the welding trajectory.

[0075] The communication interface is built with a high-speed data communication bus to realize data interaction between modules. The fault self-diagnosis module monitors the sensor data, actuator status, and communication signals of the robot in real time. When an abnormal situation is detected, it automatically triggers an alarm and takes preset safety protection measures (such as stopping the welding operation, switching to a backup control strategy, etc.) to ensure the safe and reliable operation of the system.

[0076] Figure 5 is a schematic structural diagram of the electronic device provided by the present invention, as Figure 5As shown in the figure, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communication bus 540. Among them, the processor 510, the communications interface 520, and the memory 530 complete communication with each other through the communication bus 540. The processor 510 may call the logical instructions in the memory 530 to execute the humanoid robot welding method, and the method includes: during the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determining the walking parameters of the humanoid robot at the next moment, where the change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; during the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding robot arm at the current moment, the walking parameters at the next moment, the pose of the welding robot arm at the current moment, and the preset welding trajectory, determining the pose of the welding robot arm at the next moment, where the welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robot arm at the current moment and the center line of the target weld seam; controlling the welding robot arm to weld based on the pose at the next moment.

[0077] In addition, when the logical instructions in the above-mentioned memory 530 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0078] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the humanoid robot welding method provided by each of the above methods. The method includes: during the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment, where the change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; during the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding robot arm at the current moment, the walking parameters at the next moment, the pose of the welding robot arm at the current moment, and a preset welding trajectory, determine the pose of the welding robot arm at the next moment, where the welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robot arm at the current moment and the target weld center line; based on the pose at the next moment, control the welding robot arm to perform welding.

[0079] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the humanoid robot welding method provided by each of the above methods. The method includes: during the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment, where the change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; during the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding robot arm at the current moment, the walking parameters at the next moment, the pose of the welding robot arm at the current moment, and a preset welding trajectory, determine the pose of the welding robot arm at the next moment, where the welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robot arm at the current moment and the target weld center line; based on the pose at the next moment, control the welding robot arm to perform welding.

[0080] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0081] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A humanoid robot welding method, characterized in that, Including: During the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment. The change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; During the walking and welding process of the humanoid robot, based on the welding trajectory deviation of the welding manipulator at the current moment, the walking parameters at the next moment, the pose of the welding manipulator at the current moment, and the preset welding trajectory, determine the pose of the welding manipulator at the next moment. The welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding manipulator at the current moment and the center line of the target weld; Based on the pose at the next moment, control the welding manipulator to weld.

2. The humanoid robot welding method according to claim 1, wherein The determining the walking parameters of the humanoid robot at the next moment based on the change amount of the center of gravity of the humanoid robot at the current moment includes: Obtain the motion state of the humanoid robot at the current moment; Input the motion state at the current moment and the change amount of the center of gravity at the current moment into the reinforcement learning model, and obtain the walking parameters at the next moment output by the reinforcement learning model; The reinforcement learning model is trained with the goal of minimizing the reward value, and the reward value is determined based on the change amount of the center of gravity of the humanoid robot.

3. The humanoid robot welding method according to claim 2, wherein, The determining the pose of the welding manipulator at the next moment based on the welding trajectory deviation of the welding manipulator at the current moment, the walking parameters at the next moment, the pose of the welding manipulator at the current moment, and the preset welding trajectory includes: Based on the preset welding trajectory, construct a dynamic model during the walking process of the humanoid robot. The dynamic model is used to describe the motion relationship and force situation between components during the walking process of the humanoid robot, as well as the kinematic constraints of the welding manipulator; Based on the dynamic model, the welding trajectory deviation at the current moment, the walking parameters at the next moment, and the pose of the welding manipulator at the current moment, determine the optimal adjustment path and pose compensation amount of the welding manipulator; Based on the optimal adjustment path and the pose compensation amount, determine the pose of the welding manipulator at the next moment.

4. The humanoid robot welding method according to claim 2, characterized in that, The obtaining steps of the motion state at the current moment and the pose at the current moment are as follows: Collect the initial motion state at the current moment and the initial pose at the current moment through sensors; Filter the initial motion state at the current moment to obtain the motion state at the current moment; Filter the initial pose at the current moment to obtain the pose at the current moment.

5. The humanoid robot welding method according to any one of claims 1 to 4, characterized in that, The method further includes: In the case where the change amount of the center of gravity at the current moment is greater than the first threshold, and / or the welding trajectory deviation at the current moment is greater than the second threshold, give a warning prompt.

6. A humanoid robot welding device, characterized in that, Including: A first determination unit, configured to, during the walking and welding process of the humanoid robot, based on the change amount of the center of gravity of the humanoid robot at the current moment, determine the walking parameters of the humanoid robot at the next moment. The change amount of the center of gravity at the current moment refers to the deviation between the center of gravity position at the current moment and the center of gravity position at the previous moment; A second determination unit, configured to determine the pose of the welding robotic arm at the next moment based on the welding trajectory deviation of the welding robotic arm at the current moment, the walking parameters at the next moment, the pose of the welding robotic arm at the current moment, and a preset welding trajectory during the walking and welding process of the humanoid robot, where the welding trajectory deviation at the current moment refers to the deviation between the actual welding trajectory of the welding robotic arm at the current moment and the center line of the target weld seam; A welding unit, configured to control the welding robotic arm to perform welding based on the pose at the next moment.

7. A humanoid robot, characterized in that, Comprising: A robot body, a welding robotic arm installed on the robot body, and the humanoid robot welding device according to claim 6.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, wherein, When the processor executes the computer program, the humanoid robot welding method according to any one of claims 1 to 5 is implemented.

9. A non-transitory computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, the humanoid robot welding method according to any one of claims 1 to 5 is implemented.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, the humanoid robot welding method according to any one of claims 1 to 5 is implemented.

Images