A human-robot collaborative intelligent robot motion control system and method

Abstract

This invention discloses a human-robot collaborative intelligent robot motion control system and method, relating to the field of robot control technology. The system includes using environmental position changes and robot body sway as offsetting reference values; removing environmental interference and robot inertial forces from the physical thrust values ​​at the contact surface to obtain the net thrust value of the human's active force; combining the net thrust value with limb displacement for intent analysis; generating different drives based on the intent and executing different boundary deformation commands; and releasing commands only within the boundaries. During drive execution, when the net thrust rapidly decays to zero, indicating limb disengagement and meeting safety conditions, a linearly increasing reverse resistance is generated until the robot stops. In this invention, applying reverse resistance assistance or compliant advance displacement based on the clarity of the operational intent improves collaborative stability and smoothness.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, specifically to an intelligent robot motion control system and method for human-robot collaboration. Background Technology

[0002] As the manufacturing industry evolves towards flexibility and personalization, the application scenarios of robots working collaboratively with humans are gradually increasing. Existing industrial robot motion control systems mostly adopt a hierarchical architecture. Their data acquisition mainly relies on the robot's own internal sensors (such as joint position, speed, and torque sensors) and external obstacle distance sensors. Under this architecture, the system generates control commands based on preset motion programs and trajectory parameters, and controls the motor to execute predetermined path motion through servo drives. This control mode was initially designed for independent operation scenarios isolated by physical fences. When personnel need to intervene, the conventional procedure is to first stop the robot's automatic operation mode, and after confirming the safety conditions, then input new motion commands through manual guidance or a teach pendant to complete the interaction.

[0003] In the process of direct human-machine collaboration, the physical thrust values ​​collected on the contact surface of existing robots during operation not only include the component of the force actively exerted by the human, but also the additional interference force caused by changes in the environmental position and the additional inertial force generated by the swaying of the robot body. This makes it difficult for the system to accurately extract the net thrust that reflects the true intention of the human. At the same time, during the process of the human applying the thrust, the limb movements have multiple states such as regressive or stable displacement. The system needs to accurately distinguish between the pending and clear states of the operation intention, and achieve the coordination of compliant movement and advanced displacement when the intention is clear. Summary of the Invention

[0004] To address the aforementioned technical problems, this paper provides a human-machine collaborative intelligent robot motion control system and method. This solution resolves the issue raised in the background section regarding the physical thrust values ​​collected on the robot's contact surface during direct human-machine interaction. These values ​​include not only the force exerted by the human operator but also additional forces from environmental position changes and inertial forces from the robot's own movement. This makes it difficult for the system to accurately extract the net thrust reflecting the operator's true intention. Furthermore, during the application of thrust, limb movements can exhibit various states, such as regressive or stable displacement. The system needs to accurately distinguish between the pending and clear states of intention and coordinate compliant and proactive movements when the intention is clear.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A human-machine collaborative intelligent robot motion control method, specifically including: The data collected include the displacement amount, displacement direction, and physical thrust value of the contact surface. The data also includes the environmental position change and the robot body's swaying. The environmental position change and the robot body's swaying are used as offset reference values. The environmental interference force and the robot's inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force exerted by the person. This data is then combined with the limb displacement to form comprehensive reference information. The comprehensive reference information is analyzed. When the personnel limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending. An auxiliary stabilization command to apply reverse physical resistance is generated. When the personnel limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear. A compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position. The invisible boundary that allows the robot to move is defined by combining the changes in the environmental position. When the operation intention is clear, the invisible boundary is expanded forward along the direction to accommodate the advance displacement, while the lateral boundary margin is contracted to prevent yaw. The command is only released within the invisible boundary. The release command is converted into a power signal for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

[0006] Preferably, the collected personnel limb displacement, displacement direction, and physical thrust values ​​at the contact surface are combined with the collected environmental position changes and robot body swaying. These environmental position changes and robot body swaying are used as offset reference values. The environmental interference and robot inertial forces are removed from the physical thrust values ​​at the contact surface to obtain the net thrust value generated purely by the personnel's active force. This combined with the limb displacement data forms comprehensive reference information, specifically including: Obtain the force trend in the first direction corresponding to the change in environmental position, and the force trend in the second direction corresponding to the shaking of the robot body; The force trend in the first direction is compared and canceled out with the component of the physical thrust at the contact surface that deviates from the direction of the person's limb displacement, thus eliminating the additional force caused by environmental interference. The time delay caused by the robot's shaking motion relative to the physical thrust value at the contact surface is obtained. The force trend in the second direction is aligned with the physical thrust value of the contact surface according to the time delay. The part that exceeds the thrust threshold corresponding to the stable displacement after alignment is marked as the robot's inertial additional force and then stripped away. The remaining value after offsetting the additional forces from environmental interference and stripping away the additional forces from the robot's inertia is recorded as the net thrust generated by the active force exerted by the personnel. The net thrust value and the displacement of personnel limbs are combined as comprehensive reference information.

[0007] Preferably, in the process of analyzing the comprehensive reference information, when the reciprocating movement of a person's limbs is detected and the net thrust value exhibits unstable periodic fluctuations, the operational intent is determined to be pending, and an auxiliary stabilization command to apply reverse physical resistance is generated. When the stable displacement of a person's limbs is detected and the net thrust value has a stable direction, the operational intent is determined to be clear, a compliant movement command to reduce physical resistance is generated, and an advanced displacement command is generated based on the direction to calculate the desired position. Specifically, this includes: Within a preset time window, the peak and trough values ​​of the net thrust are extracted. When the difference between the peak and trough values ​​exceeds a preset fluctuation threshold and the displacement of the human limbs reverses direction in space, it is confirmed that an unstable periodic fluctuation is present. The operation intention is determined to be pending. The instantaneous reverse direction of the human limbs' reciprocating displacement is calculated, and an auxiliary stabilization command is generated to apply physical resistance in the same direction as the instantaneous reverse direction to the robot joints. When the rate of change of velocity corresponding to the displacement of the human limb is lower than the preset stability threshold, and the angle of change of direction corresponding to the net thrust value is lower than the preset angle threshold, it is confirmed that there is a stable direction, the operation intention is clear, and a compliant movement instruction is generated to adjust the current physical resistance parameter of the robot to the compliant following mode. The net thrust value, the velocity value corresponding to the displacement of the person's limb, and the displacement direction are obtained at the moment when the intention of the operation is clear. The velocity value and the net thrust value are calculated with a preset feedforward coefficient. Combined with the displacement direction, the spatial coordinates after a preset time interval are calculated as the desired position, and an advance displacement command with the desired position as the execution target is generated.

[0008] Preferably, the step of defining an invisible boundary that allows robot movement based on environmental position changes, and when the operational intent is clear, expanding the invisible boundary forward along the direction to accommodate advance displacement, while simultaneously contracting the lateral boundary margin to prevent yaw, and issuing commands only within the invisible boundary, specifically includes: Extract the spatial coordinate offset corresponding to the environmental position change, and update the initial safe workspace by translation based on the spatial coordinate offset to generate an invisible boundary that allows the robot to move at the current moment. When the operation intention is clear, the target spatial coordinates corresponding to the advance displacement command are extracted, the forward straight-line distance between the target spatial coordinates and the robot's current spatial coordinates is calculated, and the invisible boundary is extended outward in the forward dimension corresponding to the direction equal to the length of the forward straight-line distance. Extract the velocity value corresponding to the stable displacement of the person's limbs, calculate the lateral yaw prediction deviation based on the velocity value, and reduce the width of the invisible boundary inward in the lateral dimension perpendicular to the direction to be equal to the lateral yaw prediction deviation. The system compares the expected execution coordinates of each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, a release signal is output; when the expected execution coordinates are outside the invisible boundary, the corresponding instruction is intercepted.

[0009] Preferably, the step of converting the release command into a power signal for driving, determining that the person's limb has detached when the net thrust value rapidly decreases to zero, obtaining the robot's load at the moment of detachment, and generating resistance that gradually increases in the opposite direction of the original movement when the load is below a safety threshold and there is no obstruction in front, until the robot stops moving, specifically includes: Within a preset sampling period, when the rate of change of the net thrust value exceeds the preset attenuation slope threshold and the value falls within the zero tolerance range, it is determined that the person's limb has been removed. The robot's current equivalent load at the moment the person's limb is separated is used as the robot's load, and environmental spatial data is combined to determine whether there are obstacles in front of the robot along its current trajectory. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's motion speed at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value. Using the opposite direction of the robot's current motion as the direction of resistance application, and using the initial kinetic energy reference value as the initial calculation base for resistance, a gradually increasing reverse resistance torque is generated according to a preset linearly increasing slope. The reverse resistance torque is superimposed on the power signal and continuously output to absorb the residual kinetic energy generated by the robot's operation until the robot's speed is detected to drop to zero.

[0010] Preferably, the motion control method further includes: When the robot's load is lower than a preset safe mass threshold and it is determined that there are no obstacles ahead, the robot's motion velocity at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value, specifically including: The robot's load capacity is compared with a preset safety quality threshold, which is a certain percentage of the robot's rated load capacity calibrated at the factory. The robot retrieves the surrounding spatial distance information provided by the environmental perception unit and searches for whether there are any objects within a preset angle range in front of the robot's current direction of movement. If no objects are found, it is determined that there are no obstacles in front of it. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's end-effector speed value recorded at the sampling moment before the departure determination occurs is extracted from the speed recording cache and used as the initial kinetic energy reference value.

[0011] A human-machine collaborative intelligent robot motion control system, used to execute the motion control method described above, specifically includes: The human-machine collaboration system is used to acquire the displacement amount, displacement direction and physical thrust value of the contact surface of the human limbs, and analyze the comprehensive reference information. When the human limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending, and an auxiliary stabilization command to apply the reverse physical resistance is generated. When the human limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear, a compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position.

[0012] Preferably, the motion control system further includes: The robot information monitoring and acquisition system is used to collect environmental position changes and robot body movement. The environmental position changes and robot body movement are used as offset reference values. The environmental interference force and robot inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force of the human. Combined with the limb displacement, a comprehensive reference information is formed.

[0013] Preferably, the motion control system further includes: The robot storage and verification system receives various instructions, defines the invisible boundary that allows the robot to move by combining the changes in the environmental position, and expands the invisible boundary forward along the direction to accommodate the advance displacement when the operation intention is clear. At the same time, it shrinks the lateral boundary margin to prevent yaw. It compares the expected execution coordinates corresponding to each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, it outputs a release signal. When the expected execution coordinates are outside the invisible boundary, it intercepts the corresponding instruction.

[0014] Preferably, the motion control system further includes: The robot control and drive system is used to convert release commands into power signals for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: In this invention, by using environmental position changes and robot body sway as offset references, and utilizing time delay to perform temporal alignment and threshold stripping of the force trend in the second direction, the environmental interference additional force and the robot inertial additional force in the physical thrust value of the contact surface are effectively separated, which helps to improve the accuracy of the system in obtaining the net thrust value of the active force exerted by the personnel.

[0016] The present invention uses comprehensive reference information based on net thrust value and personnel limb displacement, combined with parameters such as peak-to-trough difference, direction reversal and velocity change rate, to distinguish between the pending and clear states of operational intent, and generate corresponding auxiliary stabilization commands with reverse physical resistance or advance displacement commands containing the desired position, so that the robot's motion response state is more in line with the actual operational intent of the personnel.

[0017] In this invention, the position of the invisible boundary is updated by combining the environmental position change. When the operation intention is clear, an asymmetric deformation strategy is adopted to extend the forward dimension to accommodate the advance displacement and reduce the lateral dimension to offset the yaw prediction deviation. This not only meets the robot's mobile space requirements, but also helps to improve the stability of trajectory execution. Attached Figure Description

[0018] Figure 1 This is a flowchart of the control method in this invention; Figure 2 This is a flowchart of the comprehensive reference information acquisition steps in this invention; Figure 3 This is a flowchart of the instruction acquisition and verification steps in this invention; Figure 4 This is a diagram illustrating the operational framework of the control system in this invention. Detailed Implementation

[0019] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0020] Reference Figure 1 As shown, a human-machine collaborative intelligent robot motion control method specifically includes: The data collected include the displacement amount, displacement direction, and physical thrust value of the contact surface. The data also includes the environmental position change and the robot body's swaying. The environmental position change and the robot body's swaying are used as offset reference values. The environmental interference force and the robot's inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force exerted by the person. This data is then combined with the limb displacement to form comprehensive reference information. The comprehensive reference information is analyzed. When the personnel limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending. An auxiliary stabilization command to apply reverse physical resistance is generated. When the personnel limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear. A compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position. The invisible boundary that allows the robot to move is defined by combining the changes in the environmental position. When the operation intention is clear, the invisible boundary is expanded forward along the direction to accommodate the advance displacement, while the lateral boundary margin is contracted to prevent yaw. The command is only released within the invisible boundary. The release command is converted into a power signal for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

[0021] Furthermore, the collected data include the amount of limb displacement, displacement direction, and physical thrust at the contact surface. The data also includes the environmental position change and the robot's swaying motion. These are used as offsetting reference values. Environmental interference and robot inertia forces are removed from the physical thrust at the contact surface to obtain the net thrust value generated purely by the person's active force. This information is then combined with the limb displacement to form comprehensive reference information, specifically including: Obtain the force trend in the first direction corresponding to the change in environmental position, and the force trend in the second direction corresponding to the shaking of the robot body; The force trend in the first direction is compared and canceled out with the component of the physical thrust at the contact surface that deviates from the direction of the person's limb displacement, thus eliminating the additional force caused by environmental interference. The time delay caused by the robot's shaking motion relative to the physical thrust value at the contact surface is obtained. The force trend in the second direction is aligned with the physical thrust value of the contact surface according to the time delay. The part that exceeds the thrust threshold corresponding to the stable displacement after alignment is marked as the robot's inertial additional force and then stripped away. The remaining value after offsetting the additional forces from environmental interference and stripping away the additional forces from the robot's inertia is recorded as the net thrust generated by the active force exerted by the personnel. The net thrust value and the displacement of personnel limbs are combined as comprehensive reference information; In this scheme, the acceleration information corresponding to the change in environmental position is obtained by sensors arranged on the robot base. Combined with the mass parameters of the robot and the load, the force trend caused by the change in environmental position is calculated and recorded as the first direction force trend. The motion state of each joint is obtained by sensors installed on each joint of the robot, including the rotation angle, rotation speed and the degree of change of rotation speed. Based on the mass and center of gravity distribution characteristics of each segment of the robot's arm, the internal force trend generated by the robot's own motion at the end contact surface is calculated and recorded as the second direction force trend. The first direction force trend relies on the continuous measurement of the base’s swaying speed in the front-back, left-right, and up-down directions by the sensor. The swaying speed change value in each direction is multiplied by the total weight of the robot itself plus the weight of the item being carried. The product is the magnitude of the pushing or pulling trend caused by environmental changes in that direction. The magnitude of the trend in the three directions together with the corresponding direction constitutes the first direction force trend. The second-direction force trend relies on rotation angle measurement sensors installed at each of the robot's moving joints to acquire the rotation angle, rotation speed, and changes in rotation speed of each joint in real time. Combined with the weight of each arm segment, the distance from the arm's center of gravity to the joint axis, and the inertia characteristics of the arm rotating around the axis, which were recorded at the time the robot left the factory, the internal pushing and pulling forces generated by the movement of each joint on the arm segment are accumulated and transmitted to the end position where the robot contacts the human hand. After summarizing, the internal force trend generated by the robot's own movement at the end contact surface is obtained, which is recorded as the second-direction force trend. The direction of the personnel's limb displacement is extracted and used as the main force reference direction. The physical thrust value of the contact surface is decomposed along this reference direction to obtain the tangential component force consistent with this direction and the normal component force perpendicular to this direction. The force trend in the first direction is compared with the normal component force in terms of numerical magnitude and change rate. When the difference between the force trend in the first direction force trend and the normal component force in terms of numerical magnitude and change rate is less than the preset interference matching threshold value, it is determined that the normal component force is mainly caused by the change of environmental position. It is regarded as the environmental interference additional force and subtracted from the physical thrust value of the contact surface. The preset interference matching threshold value is obtained by data collection and statistics under static working conditions at the time of factory delivery. The time lag in the transmission of the internal forces generated by the robot's own motion to the end contact surface is obtained. The time lag can be obtained through factory testing. Based on the time lag, the sequence of the force trend in the second direction and the sequence of the physical thrust values ​​of the contact surface are translated and aligned in time. A reference value of the basic maintaining force corresponding to the stable displacement is set. The force trend in the second direction after time alignment is compared with the reference value of the basic maintaining force. Only the excess part is marked as the robot's inertial additional force, and the excess part is subtracted from the physical thrust value of the contact surface. The part that does not exceed the limit is reserved as the force required to maintain normal motion. After subtracting the environmental interference additional force and the robot's inertial additional force from the physical thrust value of the contact surface in sequence, the remaining value is recorded as the net thrust value generated by the active force exerted by the personnel. The system obtains the estimated resistance values ​​generated by internal friction of the robot at the current rotation speed of each joint. The estimated resistance values ​​are obtained by considering the rotation speed and the viscous friction coefficient, which is obtained from factory testing. The system also obtains the remaining deviation value after gravity compensation calculation of the robot at the current posture. The deviation value is obtained by converting the difference between the theoretical maintaining torque output by the robot's own gravity balance algorithm and the actual joint output force feedback value into the force at the end contact surface. During the conversion, the system uses the current rotation angle of each joint of the robot to construct the transmission relationship matrix between joint torque and end force. The elements of the matrix are updated in real time as the robot's posture changes. The vector composed of the torque differences of each joint is calculated with the transmission relationship matrix. The result is the equivalent force generated at the end contact surface by the gravity compensation residual of each joint. The fluctuation range of the end sensor reading within the sampling time is obtained when the robot is stationary and without external force contact. The fluctuation range is obtained by recording the difference between the maximum and minimum values ​​of the sensor reading within the sampling time during the factory static calibration. The magnitude of the joint friction resistance, the amplitude of the gravity compensation residual conversion value, and the noise fluctuation range of the end sensor are added to obtain a comprehensive background noise amplitude threshold, which is used as the reference value of the basic maintaining force corresponding to the stable displacement. The sampling time is not less than 30 seconds. Finally, the data on the displacement of the human limbs provided by the external motion capture device or the wearable position sensing device are obtained. Based on the time lag, the sequence of net thrust values ​​and the sequence of human limb displacement values ​​are time-labeled and aligned. The aligned net thrust values ​​and limb displacement values ​​are then merged one by one to form a comprehensive reference information group.

[0022] Furthermore, the analysis of comprehensive reference information, when detecting reciprocating movement of a person's limbs and unstable periodic fluctuations in net thrust, determines that the operational intent is pending and generates an auxiliary stabilization command to apply reverse physical resistance. When detecting stable displacement of a person's limbs and stable direction of net thrust, determines that the operational intent is clear and generates a compliant movement command to reduce physical resistance. Based on the direction, a forward displacement command is generated to calculate the desired position. Specifically, this includes: Within a preset time window, the peak and trough values ​​of the net thrust are extracted. When the difference between the peak and trough values ​​exceeds a preset fluctuation threshold and the displacement of the human limbs reverses direction in space, it is confirmed that an unstable periodic fluctuation is present. The operation intention is determined to be pending. The instantaneous reverse direction of the human limbs' reciprocating displacement is calculated, and an auxiliary stabilization command is generated to apply physical resistance in the same direction as the instantaneous reverse direction to the robot joints. When the rate of change of velocity corresponding to the displacement of the human limb is lower than the preset stability threshold, and the angle of change of direction corresponding to the net thrust value is lower than the preset angle threshold, it is confirmed that there is a stable direction, the operation intention is clear, and a compliant movement instruction is generated to adjust the current physical resistance parameter of the robot to the compliant following mode. The net thrust value, the velocity value corresponding to the displacement of the personnel limb, and the displacement direction are obtained at the moment when the intention of the operation is clear. The velocity value and the net thrust value are calculated with a preset feedforward coefficient. Combined with the displacement direction, the spatial coordinates after a preset time interval are calculated as the desired position, and an advance displacement command with the desired position as the execution target is generated. In this scheme, the highest and lowest values ​​of the net thrust are continuously extracted within a preset time window. The preset time window is long enough to fully encompass at least two potential tremor cycles, and is determined based on the range of physiological tremor frequencies of the human hand and the underlying sampling period. When the difference between the highest and lowest values ​​is greater than the preset fluctuation judgment standard value, and the direction of the human limb displacement changes in space, it is confirmed that the net thrust value exhibits unstable periodic fluctuations. The operation intention is determined to be in a pending state. The current reverse direction of the human limb reciprocating movement is calculated, and an auxiliary stabilization command is generated to apply physical resistance in the same direction as the current reverse direction to the robot's moving joints. When the change in the speed of movement corresponding to the displacement of the human limb is lower than the preset stable judgment standard value, and the angle of change of the force direction corresponding to the net thrust value is lower than the preset angle judgment standard value, it is confirmed that the net thrust value has a stable direction, the operation intention is determined to be clear, and a compliant movement instruction is generated to reduce the robot's current physical resistance parameter to a compliant following state. The two judgment logics correspond to different movement stages and force characteristics in the operation process, and the physical quantity threshold conditions they are based on are mutually exclusive and will not be satisfied at the same time. There is no problem of judgment conflict or state overlap. When the features of the comprehensive reference information do not meet the judgment conditions of pending operation intention or clear operation intention, the current control state is maintained, and the comprehensive reference information of the next time window is collected and a new round of judgment is performed until any set of judgment conditions is met. If any set of judgment conditions cannot be met in multiple consecutive preset time windows, the operation intention is judged to be in a pending state by default, and an auxiliary stabilization command is generated to apply physical resistance to the robot's active joints in the opposite direction to the most recently detected direction of human limb movement. The net thrust value, the movement speed value corresponding to the displacement of the personnel limb, and the movement direction of the personnel limb are obtained when the determination of the operation intention is clear. The movement speed value and the net thrust value are substituted into the pre-set feedforward calculation coefficient for calculation. Then, combined with the movement direction, the spatial coordinates of the personnel limb to be reached after a pre-set time length are calculated. The calculated spatial coordinates are used as the expected position to be reached, and a forward displacement command with the expected position to be reached as the execution target is generated. The preset feedforward operation coefficients are obtained as follows: the robot is placed in a free state without external load, and the operator applies a constant test thrust of known magnitude to the robot's end effector. The actual acceleration value generated by the robot's end effector under the action of the constant test thrust is recorded. According to Newton's second law, the constant test thrust is divided by the actual acceleration value to calculate the equivalent end effector mass of the robot in the current posture. The preset feedforward operation coefficients are composed of the reciprocal of the equivalent end effector mass.

[0023] Furthermore, the process of defining an invisible boundary that allows robot movement based on environmental position changes, and when the operational intent is clear, expanding the invisible boundary forward along the direction to accommodate advance displacement, while simultaneously contracting the lateral boundary margin to prevent yaw, and issuing commands only within the invisible boundary, specifically includes: Extract the spatial coordinate offset corresponding to the environmental position change, and update the initial safe workspace by translation based on the spatial coordinate offset to generate an invisible boundary that allows the robot to move at the current moment. When the operation intention is clear, the target spatial coordinates corresponding to the advance displacement command are extracted, the forward straight-line distance between the target spatial coordinates and the robot's current spatial coordinates is calculated, and the invisible boundary is extended outward in the forward dimension corresponding to the direction equal to the length of the forward straight-line distance. Extract the velocity value corresponding to the stable displacement of the person's limbs, calculate the lateral yaw prediction deviation based on the velocity value, and reduce the width of the invisible boundary inward in the lateral dimension perpendicular to the direction to be equal to the lateral yaw prediction deviation. The system compares the expected execution coordinates of each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, a pass signal is output. When the expected execution coordinates are outside the invisible boundary, the corresponding instruction is intercepted. In this scheme, the spatial coordinate offset corresponding to the change in environmental position is obtained. The spatial coordinate offset reflects the global reference position drift caused by ground micro-movement, workbench displacement, or slight slippage of robot base. The coordinates of each boundary vertex of the pre-set initial safe workspace are superimposed with the displacement components of the spatial coordinate offset in the corresponding dimension to complete the overall position translation of the initial safe workspace. The spatial area obtained after translation is the invisible boundary that allows the robot to move at the current moment. The initial safe workspace is determined by the personnel manually guiding the robot end effector to move to the farthest position allowed to be reached in the work area, and the limit coordinates of the end effector in each direction are recorded. All coordinate points together form a closed three-dimensional spatial range as the initial safe workspace. The invisible boundary is updated synchronously with the changes in the environmental position, ensuring that the safe area always maintains a relatively fixed positional relationship with the actual working environment. When the operation intention is determined to be clear, the target spatial coordinates corresponding to the advance displacement command are extracted, the Euclidean distance between the target spatial coordinates and the current spatial coordinates of the robot end effector are calculated, and the Euclidean distance is projected along the forward direction corresponding to the operation intention to obtain the forward straight-line distance. The direction vector of the operation intention is determined by the direction of the personnel's limb displacement at the moment the operation intention is determined to be clear. The updated invisible boundary is translated outward in the forward direction by the length of the forward straight-line distance, so that the invisible boundary is expanded in the forward dimension to cover the expected position range to be reached by the advance displacement command, avoiding the command being intercepted due to the advance movement distance exceeding the original boundary. Simultaneously, the movement speed value corresponding to the stable displacement of the human limb is extracted, and the lateral yaw prediction deviation is calculated based on the movement speed value. The lateral yaw prediction deviation is used to estimate the deviation magnitude that the robot may have in the forward movement process. The width of the lateral yaw prediction deviation is contracted inward from both sides of the invisible boundary in the lateral direction perpendicular to the direction of the operation intention. After the forward expansion and lateral contraction, the invisible boundary expands the forward range of motion while maintaining a safe distance in the lateral direction to prevent accidental yaw. The calculation method for the lateral yaw prediction deviation is as follows: During the robot's factory debugging phase, the robot is controlled to perform pure forward compliant following motion at different end-effector movement speeds without any operator applying lateral force. The maximum offset distance of the robot's end-effector in the direction perpendicular to the forward direction and the forward distance traveled by the robot's end-effector when the maximum offset occurs are recorded in each motion. The ratio of the maximum offset distance to the forward distance is calculated as the yaw rate at that speed. Each set of movement speed values ​​and the corresponding yaw rate are recorded as a correspondence, and a lookup table between speed values ​​and yaw rates is established. During the actual operation of the robot, when it is necessary to calculate the lateral yaw prediction deviation, the movement speed value corresponding to the current stable displacement of the operator's limb is extracted. The corresponding yaw rate is found according to the lookup table. The yaw rate is multiplied by the forward straight distance corresponding to the current forward displacement command. The product is the lateral yaw prediction deviation. The system receives various commands to be executed, including auxiliary stabilization commands, compliant movement commands, advanced displacement commands, and motion commands. For each command, it extracts the expected spatial coordinates that the robot's end effector will reach after execution. It then compares these expected spatial coordinates with the dynamically updated invisible boundary to determine their spatial inclusion relationship. If the expected spatial coordinates are within the three-dimensional space defined by the invisible boundary, it outputs a release signal, allowing the command to proceed to the subsequent drive conversion stage. If the expected spatial coordinates are outside the invisible boundary, it intercepts the corresponding command and does not allow it to proceed.

[0024] Furthermore, the process of converting the release command into a power signal for driving is described. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has detached. The robot's load at the moment of detachment is obtained. When the load is below a safety threshold and there is no obstruction in front, a resistance force is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving. Specifically, this includes: Within a preset sampling period, when the rate of change of the net thrust value exceeds the preset attenuation slope threshold and the value falls within the zero tolerance range, it is determined that the person's limb has been removed. The robot's current equivalent load at the moment the person's limb is separated is used as the robot's load, and environmental spatial data is combined to determine whether there are obstacles in front of the robot along its current trajectory. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's motion speed at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value. Using the opposite direction of the robot's current motion as the direction of resistance application, and using the initial kinetic energy reference value as the initial calculation base for resistance, a gradually increasing reverse resistance torque is generated according to a preset linearly increasing slope. The reverse resistance torque is superimposed on the power signal and continuously output to absorb the residual kinetic energy generated by the robot's operation until the robot's speed is detected to drop to zero. In this scheme, during the process of driving each moving joint of the robot, the net thrust value is continuously sampled at fixed time intervals. Within three consecutive sampling cycles, the ratio of the decrease in net thrust value compared to the previous cycle to the sampling interval is calculated to obtain the rate of change of decrease. When the rate of change of decrease is greater than the preset attenuation slope judgment standard value, and the current reading of net thrust value has entered the tolerance range of small fluctuations allowed above and below zero, it is determined that the operator's hand has been separated from the robot contact surface, triggering the generation of the braking process after separation. After separation, the equivalent load mass borne by each moving joint of the robot at the moment of separation is recorded. The equivalent load mass is obtained as follows: Based on the current rotation angle, rotation speed, and driving torque of each joint, combined with the weight distribution parameters of each segment of the robot's arm and the weight distribution parameters of each segment of the robot recorded at the factory, the theoretical self-load torque required to maintain the robot's own movement and posture in the current motion state is calculated. The actual output driving torque of each joint is subtracted from the corresponding theoretical self-load torque to obtain the residual torque generated by each joint due to bearing the external load. Based on the robot's spatial posture determined by the current rotation angle of each joint, the residual torque of each joint is converted and summarized to the end position of the robot according to the torque transmission relationship of the robotic arm, so as to obtain the total external force value borne by the end. At the moment when the human limb is determined to be detached, the total external force is mainly caused by the gravity of the object being transported. The magnitude of the component of the total external force in the vertical downward direction is extracted and divided by the gravitational acceleration constant. The obtained value is the external weight borne by the end of the robot. This external weight value is used as the robot's load. The method for obtaining the attenuation slope judgment standard value is as follows: the operator simulates the normal release action of leaving the contact surface multiple times, records the process curve of the net thrust value decreasing from a stable value to close to zero during each release, calculates the decrease of force value per unit time during each release, and obtains a set of data on the rate of decrease corresponding to normal release. At the same time, the change process of force value is collected under non-real release situations such as the operator accidentally touching the contact surface slightly and then quickly removing it or the brief force fluctuation caused by hand slippage. The distribution range of the rate of decrease of force value under the two situations of normal release and non-real release is compared, and a value between the two is selected as the attenuation slope judgment standard value. When the robot's load is less than 75% of the rated load and it is determined that there are no obstacles blocking the forward direction, the robot's end-effector speed recorded at the sampling moment before the separation occurs is extracted. The end-effector speed is used as the initial kinetic energy reference value. The opposite direction of the robot's current movement direction is used as the direction of resistance application. The initial kinetic energy reference value is used as the starting base for resistance calculation. According to the preset slope value that increases linearly with time, the target value of the reverse resistance torque is gradually increased in each sampling period after separation. The linear increase slope of the reverse resistance torque is pre-stored at the factory. Its value is such that the braking process is neither too abrupt to cause mechanical shock, nor too slow to cause excessive stopping distance. When the robot's load exceeds 75% of the rated load and it is determined that there are no obstacles blocking the forward direction, an immediate stop command is sent to the drive components of each joint to interrupt the continuous output of the power signal. At the same time, the mechanical brake device installed at the joint pivot is activated to apply a locking force to the joint pivot and prevent the joint from continuing to rotate. During the stopping process, the status information of the overload and the corresponding time stamp are recorded together for later reference. If an obstacle blocks the way forward, an immediate stop command is sent to each joint drive component to cut off the continuous output of the power signal. At the same time, the mechanical brake device is activated to lock the joint shaft, so that the robot stops moving in the shortest possible time. During the stopping process, the status information of the obstacle in front is recorded and stored for later review. The target value of the reverse resistance torque calculated in each cycle is superimposed on the power signal driving the robot joint, and a torque command opposite to the direction of movement is continuously output to the joint driving component. During the continuous action of the reverse torque, the robot's movement speed gradually decreases, and the residual kinetic energy is continuously absorbed by the reverse torque. Compared with instantaneous emergency stop, it can reduce mechanical impact, and compared with free sliding deceleration, it can shorten the braking distance. The robot continuously monitors the movement speed of its end effector. When the movement speed drops to near zero within a preset static judgment range, the increasing output of the reverse torque is stopped, and the reverse torque is maintained at its final value or switched to static holding torque, causing the robot to stop moving.

[0025] Furthermore, the motion control method also includes: When the robot's load is lower than a preset safe mass threshold and it is determined that there are no obstacles ahead, the robot's motion velocity at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value, specifically including: The robot's load capacity is compared with a preset safety quality threshold, which is a certain percentage of the robot's rated load capacity calibrated at the factory. The robot retrieves the surrounding spatial distance information provided by the environmental perception unit and searches for whether there are any objects within a preset angle range in front of the robot's current direction of movement. If no objects are found, it is determined that there are no obstacles in front of it. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's end-effector movement speed value recorded at the sampling moment before the departure determination occurs is extracted from the speed recording cache and used as the initial kinetic energy reference value. In this solution, after determining that a person's limb has been separated, the current robot load is first compared with a preset safety quality threshold, which is set to 75% of the robot's rated load to ensure the robot's safe operation. The system retrieves the surrounding spatial distance information in the robot's forward direction at the current moment. The system continuously acquires the spatial distance between surrounding objects and the robot's end effector through a distance detection device installed above the robot body or work area. A cone-shaped detection area with a preset angle range is defined in front of the robot's current direction of movement. Within this detection area, the system searches for objects whose distance from the robot's end effector is less than a preset safety distance. If the distance of all objects within the entire cone-shaped detection area is greater than or equal to the preset safety distance, it is determined that there are no obstacles blocking the forward direction. The preset angle range is set based on the following: taking the current velocity direction vector of the robot end as the central axis, it extends to both sides by a certain angle. The extension angle is determined based on the lateral offset that may occur when the robot brakes in an emergency. The preset safety distance is derived based on the shortest braking distance required for the robot to perform incremental reverse braking at the current movement speed. When the robot's load and obstacle detection conditions are met simultaneously, i.e. the robot's load is less than the preset safe mass threshold and there are no obstacles blocking the forward direction, the robot's end-effector velocity value stored in the previous sampling period before the determination of the separation time is extracted from the internal velocity recording cache. The velocity recording cache continuously stores the end-effector velocity obtained and converted by the rotation angle measurement device of each joint at fixed time intervals. Each time it is stored, the velocity data of the most recent three sampling periods are retained. The extracted end-effector velocity value is the initial kinetic energy reference value.

[0026] A human-machine collaborative intelligent robot motion control system, used to implement the motion control method described above, the motion control system comprising: The human-machine collaboration system is used to acquire the displacement amount, displacement direction and physical thrust value of the contact surface of the human limbs, and analyze the comprehensive reference information. When the human limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending, and an auxiliary stabilization command to apply the reverse physical resistance is generated. When the human limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear, a compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position.

[0027] Furthermore, the motion control system also includes: The robot information monitoring and acquisition system is used to collect environmental position changes and robot body movement. The environmental position changes and robot body movement are used as offset reference values. The environmental interference force and robot inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force of the human. Combined with the limb displacement, a comprehensive reference information is formed.

[0028] Furthermore, the motion control system also includes: The robot storage and verification system receives various instructions, defines the invisible boundary that allows the robot to move by combining the changes in the environmental position, and expands the invisible boundary forward along the direction to accommodate the advance displacement when the operation intention is clear. At the same time, it shrinks the lateral boundary margin to prevent yaw. It compares the expected execution coordinates corresponding to each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, it outputs a release signal. When the expected execution coordinates are outside the invisible boundary, it intercepts the corresponding instruction.

[0029] Furthermore, the motion control system also includes: The robot control and drive system is used to convert release commands into power signals for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A human-robot collaborative intelligent robot motion control method, characterized in that, include: The data collected include the displacement amount, displacement direction, and physical thrust value of the contact surface. The data also includes the environmental position change and the robot body's swaying. The environmental position change and the robot body's swaying are used as offset reference values. The environmental interference force and the robot's inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force exerted by the person. This data is then combined with the limb displacement to form comprehensive reference information. The comprehensive reference information is analyzed. When the personnel limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending. An auxiliary stabilization command to apply reverse physical resistance is generated. When the personnel limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear. A compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position. The invisible boundary that allows the robot to move is defined by combining the changes in the environmental position. When the operation intention is clear, the invisible boundary is expanded forward along the direction to accommodate the advance displacement, while the lateral boundary margin is contracted to prevent yaw. The command is only released within the invisible boundary. The release command is converted into a power signal for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

2. The human-robot collaborative intelligent robot motion control method according to claim 1, wherein, The collected data include the displacement amount and direction of the human limbs, the physical thrust value at the contact surface, the environmental position change, and the robot's sway. These are used as offsetting reference values. Environmental interference and robot inertia forces are removed from the physical thrust value at the contact surface to obtain the net thrust value generated purely by the human's active force. This data, combined with the limb displacement, forms comprehensive reference information, specifically including: Obtain the force trend in the first direction corresponding to the change in environmental position, and the force trend in the second direction corresponding to the shaking of the robot body; The force trend in the first direction is compared and canceled out with the component of the physical thrust at the contact surface that deviates from the direction of the person's limb displacement, thus eliminating the additional force caused by environmental interference. The time delay caused by the robot's shaking relative to the physical thrust value at the contact surface is obtained. The force trend in the second direction is aligned with the physical thrust value of the contact surface according to the time delay. The part that exceeds the thrust threshold corresponding to the stable displacement after alignment is marked as the robot's inertial additional force and then stripped away. The remaining value after offsetting the additional forces from environmental interference and stripping away the additional forces from the robot's inertia is recorded as the net thrust generated by the active force exerted by the personnel. The net thrust value and the displacement of personnel limbs are combined as comprehensive reference information.

3. The human-robot collaborative intelligent robot motion control method according to claim 2, wherein, The process of analyzing comprehensive reference information involves determining the operational intent as pending when a person's limbs are detected to be moving in the opposite direction and the net thrust value exhibits unstable periodic fluctuations. This results in an auxiliary stabilization command to apply reverse physical resistance. Conversely, when a person's limbs are detected to be moving smoothly and the net thrust value has a stable direction, the operational intent is determined to be clear. This results in a compliant movement command to reduce physical resistance and a forward displacement command to be generated based on the direction of the movement, calculated to determine the desired position. Specifically, this includes: Within a preset time window, the peak and trough values ​​of the net thrust are extracted. When the difference between the peak and trough values ​​exceeds a preset fluctuation threshold and the displacement of the human limbs reverses direction in space, it is confirmed that an unstable periodic fluctuation is present. The operation intention is determined to be pending. The instantaneous reverse direction of the human limbs' reciprocating displacement is calculated, and an auxiliary stabilization command is generated to apply physical resistance in the same direction as the instantaneous reverse direction to the robot joints. When the rate of change of velocity corresponding to the displacement of the human limb is lower than the preset stability threshold, and the angle of change of direction corresponding to the net thrust value is lower than the preset angle threshold, it is confirmed that there is a stable direction, the operation intention is clear, and a compliant movement instruction is generated to adjust the current physical resistance parameter of the robot to the compliant following mode. The net thrust value, the velocity value corresponding to the displacement of the person's limb, and the displacement direction are obtained at the moment when the intention of the operation is clear. The velocity value and the net thrust value are calculated with a preset feedforward coefficient. Combined with the displacement direction, the spatial coordinates after a preset time interval are calculated as the desired position, and an advance displacement command with the desired position as the execution target is generated.

4. The human-robot collaborative intelligent robot motion control method according to claim 3, wherein, The process involves defining an invisible boundary that allows robot movement based on changes in environmental position. When the operational intent is clear, the invisible boundary is expanded forward along the direction to accommodate advance displacement, while the lateral boundary margin is contracted to prevent yaw. Commands are only issued within the invisible boundary. Specifically, this includes: Extract the spatial coordinate offset corresponding to the environmental position change, and update the initial safe workspace by translation based on the spatial coordinate offset to generate an invisible boundary that allows the robot to move at the current moment. When the operation intention is clear, the target spatial coordinates corresponding to the advance displacement command are extracted, the forward straight-line distance between the target spatial coordinates and the robot's current spatial coordinates is calculated, and the invisible boundary is extended outward in the forward dimension corresponding to the direction equal to the length of the forward straight-line distance. Extract the velocity value corresponding to the stable displacement of the person's limbs, calculate the lateral yaw prediction deviation based on the velocity value, and reduce the width of the invisible boundary inward in the lateral dimension perpendicular to the direction to be equal to the lateral yaw prediction deviation. The system compares the expected execution coordinates of each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, a release signal is output; when the expected execution coordinates are outside the invisible boundary, the corresponding instruction is intercepted.

5. The human-robot collaborative intelligent robot motion control method according to claim 4, characterized in that, The process involves converting the release command into a power signal for driving. When the net thrust value rapidly decreases to zero, it is determined that the person's limb has detached. The robot's load at the moment of detachment is obtained. When the load is below a safety threshold and there is no obstruction in front, a resistance force is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving. Specifically, this includes: Within a preset sampling period, when the rate of change of the net thrust value exceeds the preset attenuation slope threshold and the value falls within the zero tolerance range, it is determined that the person's limb has been removed. The robot's current equivalent load at the moment the person's limb is separated is used as the robot's load, and environmental spatial data is combined to determine whether there are obstacles in front of the robot along its current trajectory. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's motion speed at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value. Using the opposite direction of the robot's current motion as the direction of resistance application, and using the initial kinetic energy reference value as the initial calculation base for resistance, a gradually increasing reverse resistance torque is generated according to a preset linearly increasing slope. The reverse resistance torque is superimposed on the power signal and continuously output to absorb the residual kinetic energy generated by the robot's operation until the robot's speed is detected to drop to zero.

6. The human-robot collaborative intelligent robot motion control method according to claim 5, wherein, When the robot's load is lower than a preset safe mass threshold and it is determined that there are no obstacles ahead, the robot's motion velocity at the time of departure from the previous sampling moment is extracted as the initial kinetic energy reference value, specifically including: The robot's load capacity is compared with a preset safety quality threshold, which is a certain percentage of the robot's rated load capacity calibrated at the factory. The robot retrieves the surrounding spatial distance information provided by the environmental perception unit and searches for whether there are any objects within a preset angle range in front of the robot's current direction of movement. If no objects are found, it is determined that there are no obstacles in front of it. When the robot's load is lower than the preset safety mass threshold and it is determined that there are no obstacles in front, the robot's end-effector speed value recorded at the sampling moment before the departure determination occurs is extracted from the speed recording cache and used as the initial kinetic energy reference value.

7. A human-robot collaborative intelligent robot motion control system for implementing the motion control method of any one of claims 1-6, characterized in that, include: The human-machine collaboration system is used to acquire the displacement amount, displacement direction and physical thrust value of the contact surface of the human limbs, and analyze the comprehensive reference information. When the human limbs are detected to move back and forth and the net thrust value shows an unstable periodic fluctuation, the operation intention is determined to be pending, and an auxiliary stabilization command to apply the reverse physical resistance is generated. When the human limbs are detected to move smoothly and the net thrust value has a stable direction, the operation intention is determined to be clear, a compliant movement command to reduce physical resistance is generated, and an advance displacement command is generated based on the direction to calculate the desired position.

8. The human-robot collaborative intelligent robot motion control system of claim 7, wherein, The motion control system further includes: The robot information monitoring and acquisition system is used to collect environmental position changes and robot body movement. The environmental position changes and robot body movement are used as offset reference values. The environmental interference force and robot inertial force are removed from the physical thrust value of the contact surface to obtain the net thrust value generated purely by the active force of the human. Combined with the limb displacement, a comprehensive reference information is formed.

9. The human-robot collaborative intelligent robot motion control system of claim 8, wherein, The motion control system further includes: The robot storage and verification system receives various instructions, defines the invisible boundary that allows the robot to move by combining the changes in the environmental position, and expands the invisible boundary forward along the direction to accommodate the advance displacement when the operation intention is clear. At the same time, it shrinks the lateral boundary margin to prevent yaw. It compares the expected execution coordinates corresponding to each received instruction with the updated invisible boundary in real time. When the expected execution coordinates are within the invisible boundary, it outputs a release signal. When the expected execution coordinates are outside the invisible boundary, it intercepts the corresponding instruction.

10. The human-robot collaborative intelligent robot motion control system of claim 9, wherein, The motion control system further includes: The robot control and drive system is used to convert release commands into power signals for driving. When the net thrust value is detected to rapidly decrease to zero, it is determined that the person's limb has been separated. The robot load at the moment of separation is obtained. When the load is lower than the safety threshold and there is no obstruction in front, a resistance is generated that is opposite to the original direction of movement and gradually increases until the robot stops moving.

Images