Man-machine cooperation safe assembly method based on safe skin of industrial robot

Abstract

The invention discloses a man-machine cooperation safe assembly method based on industrial robot safe skin. The method specifically comprises the following steps that a robot industrial environment under a man-machine cooperation assembly task is built; the robot working environment comprises an industrial robot, safe skin, an industrial personal computer and an operator; actual poses of the industrial robot and the safe skin are measured in real time, and a Kriging response model is constructed; solving the Kriging response model, and predicting the overall curved surface form of the safe skin; and a safe skin protection strategy is designed, a corresponding robot control strategy is selected according to the obtained curved surface form, the motion track of the robot is controlled, and safe proceeding of a man-machine cooperation assembly task is achieved. According to the method provided by the invention, the interaction safety during man-machine cooperation execution of the assembly task can be ensured, so that the man-machine cooperation safety level and cooperation efficiency in an aerospace large component assembly scene can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of large-load assembly of aerospace robots, and particularly relates to a human-machine collaborative safety assembly method based on an industrial robot safety skin. Background Art

[0002] In the working environment of large-load assembly operations in aerospace such as aircraft flap assembly, traditional manual hoisting operations face problems such as low efficiency, easy damage to mechanical structures, and personnel being in a highly dangerous working environment for a long time. With the continuous development of assembly technology, industrial robot assembly systems have increasingly replaced traditional hoisting and are applied in large-load assembly operations in aerospace.

[0003] However, existing industrial robots are limited by load capacity and end accuracy and are difficult to be directly applied to the assembly of ultra-large components such as spacecraft sections. Although collaborative robots have force control functions, their force sensing range cannot meet the micro-force control requirements of large-load components. Traditional safety protection devices (such as fences and laser scanning) have monitoring blind spots in dynamic assembly scenarios, cannot real-time identify sudden changes in contact forces caused by component deformation, and are even less capable of active avoidance before collision and compliant control during contact. Summary of the Invention

[0004] Aiming at the problems existing in the above-mentioned prior art, the present invention proposes a human-machine collaborative safety assembly method based on an industrial robot safety skin, aiming to significantly improve the human-machine coexistence safety level and collaborative efficiency in the assembly scenario of large aerospace components while ensuring the safety of human-machine physical interaction.

[0005] To achieve the above technical objectives, the present invention provides the following technical solutions:

[0006] A human-machine collaborative safety assembly method based on an industrial robot safety skin, specifically including:

[0007] S1. Build a robot industrial environment for human-machine collaborative assembly tasks; the robot working environment includes: an industrial robot, a safety skin, an industrial control computer, and an operator;

[0008] S2. Real-time measure the actual poses of the industrial robot and the safety skin, and construct a Kriging response model;

[0009] S3. Solve the Kriging response model to predict the curved surface shape of the safety skin;

[0010] S4. Design a safety skin protection strategy, select a corresponding robot control strategy according to the obtained curved surface shape, control the robot motion trajectory, and realize the safe progress of the human-machine collaborative assembly task.

[0011] Further, in step S1,

[0012] The industrial robot is a heavy-duty industrial robot, which carries an end effector to accept external instructions and perform a human-robot collaborative assembly task;

[0013] The safety skin is wrapped around the outside of the industrial robot, and a limited number of contact distance sensors are randomly and dispersedly arranged on its inner surface to measure the pose data and force deformation of the safety skin; the number of positions where sensors are arranged is the same as the number of positions where sensors are not arranged;

[0014] The operator is always arranged within the working range of the industrial robot, that is, there is a risk of collision between the operator and the industrial robot;

[0015] The industrial control computer is connected to the industrial robot and the safety skin through a network cable, and there is a communication connection between them. The industrial robot moves according to the point position information sent by the industrial control computer, and the industrial control computer receives the pose data sent in real time by the safety skin and the industrial robot to perceive the overall position of the system.

[0016] Further, step S2 specifically includes:

[0017] S21. Taking the base coordinate system of the industrial robot as the world coordinate system, according to the robot theoretical digital model and the actual installation position of the safety skin, record the initial theoretical poses of each contact distance sensor and the positions where no contact distance sensors are arranged on the inner surface of the safety skin in the world coordinate system;

[0018] The initial theoretical pose of the i-th contact distance sensor is expressed as:

[0019] [x i ,y i ,z i ,a i ,b i ,c i , i = 1, 2, ···, N;

[0020] Among them, x i ,y i ,z i is the initial three-dimensional coordinate of the i-th contact distance sensor, a i ,b i ,c i is the initial three-axis rotation angle of the i-th contact distance sensor, and N is the total number of contact distance sensors;

[0021] The initial theoretical pose of the t-th position where no contact distance sensor is arranged is expressed as:

[0022]

[0023] Among them, is the initial three-dimensional coordinate of the t-th position where no contact distance sensor is arranged, is the initial three-axis rotation angle at the t-th unconfigured contact distance sensor;

[0024] S22. During the operation of the safety skin, if the safety skin is deformed by external force, the contact distance sensors arranged on its inner surface are activated to obtain the actual measured pose at each sensor position; then, the deviation between the actual measured pose and the initial theoretical pose is calculated, and the formula is expressed as:

[0025]

[0026] where, are respectively the actual measured pose, the initial theoretical pose, and the deviation between the actual measured pose and the initial theoretical pose of the contact distance sensor i;

[0027] S23. Taking the positions on the inner surface of the safety skin where the contact distance sensors are not arranged as prediction points, a Kriging response model is constructed to predict the surface deformation state of the entire safety skin; the benchmark input of the Kriging response model is the deviation between the actual measured pose and the initial theoretical pose of the contact distance sensor, and the output response value is used as the pose estimation deviation at the prediction point; the Kriging response model is expressed as a combination of a regression function and a stochastic process function, and the formula is expressed as:

[0028]

[0029] where, f i (t) is the specified regression function, β i is the coefficient of the regression function, and z(t) is a stochastic process function; is the response value, that is, the estimated pose deviation;

[0030] S24. During the estimation process, to ensure the credibility of the estimation, the covariance is introduced as the evaluation criterion for the pose estimation error, and the stochastic process function is set to be a normal distribution function with a mean of 0 and a variance of σ 2 , then the covariance between adjacent prediction points t1 and t2 is expressed as:

[0031] Cov[z(t1), z(t2)] = σ 2 R(θ R , t1, t2);

[0032] where, z(t1) and z(t2) are the stochastic process functions at the prediction points t1 and t2; R(θ R , t1, t2) is the Gaussian correlation function between adjacent prediction points t1 and t2; the covariance between adjacent prediction points is used as the evaluation criterion for the prediction accuracy, and the smaller the covariance, the higher the estimation accuracy of the model; the Gaussian correlation function is specifically expressed as:

[0033]

[0034] Among them, for two adjacent predicted points t1 and t2, and are respectively the components of the estimated pose deviation at t1 and t2 in the k-th degree of freedom.

[0035] Furthermore, step S3 specifically includes:

[0036] S31. Solve the Kriging model to obtain the estimated pose deviation of the predicted point as:

[0037]

[0038] where ω T is the weighting coefficient; ε is the measurement error of the Kriging response model; ζ is the matrix composed of the pose deviations of each contact distance sensor; the weighting coefficient is obtained by constructing and solving the Lagrangian function;

[0039] S32. Denote the estimated pose deviation of any predicted point t solved by the Kriging model as:

[0040]

[0041] According to the estimated deviation of the predicted point and the initial theoretical pose, calculate the estimated pose representation of the predicted point t of the safety skin at the current moment in the world coordinate system That is, it satisfies: t = 1, 2, ···, N; where is the estimated three-dimensional coordinate of the predicted point t, is the estimated three-axis rotation angle of the predicted point t.

[0042] Furthermore, step S4 specifically includes:

[0043] S41. Establish communication between the industrial robot, the safety skin and the industrial control computer;

[0044] S42. Start the external automatic operation of the industrial robot, execute the assembly task and record the motion trajectory;

[0045] S43. Activate the contact distance sensors arranged on the inner surface of the safety skin; predict the surface deformation of the safety skin at each time step during the external automatic operation of the industrial robot, and formulate a hierarchical control strategy for the robot on the premise of ensuring the safety of the operator until the assembly task is completed.

[0046] More specifically, step S41 is specifically:

[0047] Through the ADS communication protocol, establish communication between the upper and lower computers in the industrial control computer; through the TCP communication protocol, establish communication between the lower computer and the industrial robot, and the upper computer sends instructions to the robot via the lower computer; through the UDP communication protocol and the Kuka industrial robot RSI plug-in package, establish communication between the upper computer and the industrial robot, which is used for the robot to transmit real-time positions to the upper computer and the upper computer to send motion trajectories to the robot; through the TCP communication protocol, establish communication between the lower computer and the safety skin, which is used for the safety skin to transmit contact trigger signals and distances to the lower computer.

[0048] More specifically, step S42 is specifically as follows:

[0049] The operator sends an external automatic operation instruction through the upper computer UI interface in the industrial control computer to start the external automatic operation of the robot; after the external automatic operation is started, the industrial robot continuously transmits real-time positions to the upper computer, and the upper computer continuously stores the real-time trajectories of the robot within a fixed time window; and through the Kuka industrial robot RSI plug-in package, send a target position to the robot at a fixed time step interval every time to control the robot to perform the assembly task and record the motion trajectory of the industrial robot.

[0050] More specifically, step S43 specifically includes:

[0051] S431. After activating the contact distance sensor, the robot returns to the manually set safe Home point and determines that the safe Home point is the zero position where the safety skin is not affected by external forces; the control system performs initial calibration on each contact distance sensor and calibrates the contact distance shown by each sensor to L0;

[0052] S432. According to the real-time readings of the limited contact distance sensors and the estimated poses at the predicted points of the safety skin obtained according to the Kriging response model, determine the surface deformation of the safety skin; when the safety skin undergoes surface deformation and the contact distance is less than L1, trigger the robot to decelerate;

[0053] S433. During the deceleration process of the industrial robot, continuously perform pose estimation at the predicted points and pose measurement at the sensors. When the safety skin undergoes surface deformation and the contact distance is less than L2, trigger the robot to stop urgently, and at the same time send a warning signal to the operator and the plant safety personnel;

[0054] S434. After triggering the robot to stop urgently, continuously judge the contact distance. If the contact distances at all positions return to L0, it is considered that the human and the machine are separated. After the operator confirms the separation of the human and the machine in the upper computer interface, restart the external automatic operation of the robot; if it still does not return to L0 within the set emergency stop time threshold, guide the robot to retreat through the trajectory within the fixed time window stored in the upper computer;

[0055] S435. Determine the contact distance at all times during the entire assembly task, and repeat steps S432 - S434 until the entire human - robot collaborative assembly task is finally completed.

[0056] Based on the above - mentioned technical solution, the present invention has at least the following beneficial effects:

[0057] 1. The present invention constructs a Kriging response model through the measured values of a limited number of and a small number of contact - distance sensors arranged on the inner surface of the safety skin wrapped outside the robot. While saving resource consumption, according to the real - time data of the sensors and the predicted data of the Kriging response model, a mathematical model of the safety - skin surface is established to accurately predict the surface deformation of the safety skin covering the surface of the industrial robot in real time.

[0058] 2. The present invention realizes millisecond - level motion - trajectory control and real - time pose feedback by integrating multi - protocol hierarchical communication (ADS / TCP / UDP) with the Kuka robot RSI software package; on this basis, a hierarchical safety mechanism for the safety skin is established, and combined with a trajectory - memory rollback mechanism with a fixed - time window, a three - level protection for dynamic contact distance is realized.

[0059] 3. The method proposed by the present invention not only ensures the safety of human - robot physical interaction, but also significantly improves the human - machine co - existence safety level and collaboration efficiency in the assembly scenario of large aerospace components. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic diagram of the industrial environment of the robot under the human - robot collaborative assembly task built in the method proposed by the present invention;

[0061] Figure 2 It is a flowchart of a human - robot collaborative safety assembly method based on the safety skin of an industrial robot proposed by the present invention;

[0062] Figure 3 It is a schematic diagram of the structure of the array - type contact - distance sensor in the method proposed by the present invention;

[0063] Figure 4 It is a schematic diagram of the contact distance during the whole process of sensor triggering in the method proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the above - mentioned objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the attached Figures 1-4 drawings and specific embodiments. Through this, the implementation process of how this application uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0065] Those of ordinary skill in the art can understand that all or part of the steps in implementing the method of the above embodiments can be completed by instructing relevant hardware through a program. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0066] As Figure 2 shown, the present invention proposes a human-robot collaborative safety assembly method based on the safety skin of an industrial robot, which specifically includes the following steps:

[0067] S1. Build a robot industrial environment under the human-robot collaborative assembly task; As Figure 1 shown, the robot working environment includes: an industrial robot, a safety skin, an industrial control computer, and an operator;

[0068] As a preferred embodiment, in this embodiment:

[0069] The industrial robot is a heavy-duty industrial robot, which carries an end effector to receive external instructions to execute the human-robot collaborative assembly task;

[0070] The safety skin is wrapped on the outside of the industrial robot, and a limited number of contact distance sensors are randomly and dispersedly arranged on its inner surface to measure the pose data and force deformation of the safety skin, as Figure 3 shown; the number of positions where sensors are arranged is the same as the number of positions where sensors are not arranged;

[0071] The operator is always arranged within the working range of the industrial robot, that is, it is set that there is a risk of collision between the operator and the industrial robot;

[0072] The industrial control computer is connected to the industrial robot and the safety skin through a network cable and has a communication connection with both. The industrial robot moves according to the point position information sent by the industrial control computer, and the industrial control computer receives the pose data sent by the safety skin and the industrial robot in real time to perceive the overall position of the system.

[0073] S2. Measure the actual poses of the industrial robot and the safety skin in real time and build a Kriging response model;

[0074] As a preferred embodiment, step S2 specifically includes:

[0075] S21. Take the base coordinate system of the industrial robot as the world coordinate system, and record the initial theoretical poses of the inner surface of the safety skin at each contact distance sensor and the positions where no contact distance sensors are arranged in the world coordinate system according to the robot theoretical digital model and the actual installation position of the safety skin;

[0076] The initial theoretical pose of the i-th contact distance sensor is expressed as:

[0077] [x i ,y i ,z i ,a i ,b i ,c i , i = 1, 2, ···, N;

[0078] Among them, x i ,y i ,z i are the initial three-dimensional coordinates of the i-th contact distance sensor, a i ,b i ,c i are the initial three-axis rotation angles of the i-th contact distance sensor, and N is the total number of contact distance sensors;

[0079] The initial theoretical pose at the location of the t-th unlaid contact distance sensor is expressed as:

[0080]

[0081] Among them, are the initial three-dimensional coordinates at the location of the t-th unlaid contact distance sensor, are the initial three-axis rotation angles at the location of the t-th unlaid contact distance sensor;

[0082] S22. During the operation of the safety skin, if the safety skin is deformed by external force, the contact distance sensors arranged on its inner surface are activated to obtain the actual measured pose at each sensor position; then, the deviation between the actual measured pose and the initial theoretical pose is calculated, and the formula is expressed as:

[0083]

[0084] Among them, are respectively the actual measured pose, the initial theoretical pose, and the deviation between the actual measured pose and the initial theoretical pose of the contact distance sensor i;

[0085] S23. Taking the locations on the inner surface of the safety skin where no contact distance sensors are arranged as prediction points, a Kriging response model is constructed to predict the surface deformation state of the entire safety skin; the benchmark input of the Kriging response model is the deviation between the actual measured pose and the initial theoretical pose of the contact distance sensors, and the output response value is used as the pose estimation deviation at the prediction points; in this embodiment, the Kriging response model is expressed as a combination of a regression function and a stochastic process function, and the formula is expressed as:

[0086]

[0087] Among them, f i (t) is the specified regression function, β i is the coefficient of the regression function, and z(t) is a stochastic process function; when f i (t), β i , and z(t) are known, for each prediction point t, the response function value can be solved It should be noted here that in this application, f i (t) specifically represents the combination of the deviations measured by each sensor and the initial theoretical pose at the prediction point t. By adding β i and summing to obtain the simulated surface, and then adding the stochastic process function to this surface to obtain the estimated pose deviation at the prediction point t

[0088] S24. During the estimation process, to ensure the credibility of the estimation, the covariance is introduced as the evaluation criterion for the pose estimation error, and the stochastic process function is made to be a normal distribution function satisfying a mean of 0 and a variance of σ 2 (in fact, the stochastic process errors of each prediction point are converted to the standard normal distribution here), then the covariance between adjacent prediction points t1 and t2 is expressed as:

[0089] Cov[z(t1), z(t2)] = σ 2 R(θ R , t1, t2);

[0090] Among them, z(t1) and z(t2) are the stochastic process functions at the prediction points t1 and t2; R(θ R , t1, t2) is the Gaussian correlation function between two adjacent prediction points t1 and t2; the covariance between adjacent prediction points is used as the evaluation criterion for prediction accuracy, and the smaller the covariance, the higher the estimation accuracy of the model; the Gaussian correlation function is specifically expressed as:

[0091]

[0092] Among them, for two adjacent prediction points t1 and t2, and are the components of the estimated pose deviation at t1 and t2 in the k-th degree of freedom respectively; in this application, the six-degree-of-freedom deviations are calculated separately.

[0093] S3. Solve the Kriging response model to predict the surface shape of the safety skin;

[0094] As a preferred implementation, step S3 specifically includes:

[0095] S31. Solve the Kriging model to obtain the estimated pose deviation of the prediction point as:

[0096]

[0097] Among them, ω T is the weighting coefficient, which is a matrix; ε is the measurement error of the Kriging response model; ζ is the matrix composed of the pose deviations of each contact distance sensor; the weighting coefficient is obtained by constructing and solving the Lagrangian function; in this embodiment, the formula of the Lagrangian function is expressed as:

[0098]

[0099] Among them, R is the inherent correlation matrix of the Kriging response model, which is a symmetric matrix; is the transpose of the weighting coefficient; represents the matrix of the variance of the stochastic process function;

[0100] r is a parameter only related to the deviation between the theoretical pose and the actual pose of the contact distance sensor, r = R -1 (ε - β), where β is the regression function coefficient, is its estimated value;

[0101] S32. Denote the estimated pose deviation at any prediction point t obtained by solving the Kriging model as:

[0102]

[0103] According to the estimated deviation of the prediction point and the initial theoretical pose, calculate the estimated pose representation at the safe skin prediction point t in the world coordinate system at the current moment That is, it satisfies: t = 1, 2, ···, N; where, is the estimated three-dimensional coordinate at the prediction point t, is the estimated three-axis rotation angle at the prediction point t.

[0104] So far, in the case of distance change and force deformation, the actual measured pose at the sensor position and the estimated pose at the position where the sensor is not deployed in this application are combined to obtain the surface deformation of the entire safe skin; after the safe skin is gridded according to the position, the indication of each position is displayed on the industrial control computer (as shown in (c) of Figure 4 ) as the judgment standard for the subsequent contact distance.

[0105] S4. Design a safe skin protection strategy, select the corresponding robot control strategy according to the obtained surface shape, and control the robot motion trajectory to achieve the safe progress of the human-robot collaborative assembly task;

[0106] As a preferred implementation manner, step S4 specifically includes:

[0107] S41. Establish communication between the industrial robot, the safety skin, and the industrial control computer;

[0108] In this embodiment, step S41 is specifically as follows:

[0109] Establish communication between the upper and lower computers in the industrial control computer through the ADS communication protocol; establish communication between the lower computer and the industrial robot through the TCP communication protocol, and the upper computer sends instructions to the robot via the lower computer; establish communication between the upper computer and the industrial robot through the UDP communication protocol and the Kuka industrial robot RSI plugin package, which is used for the robot to transmit real-time positions to the upper computer and the upper computer to send motion trajectories to the robot; establish communication between the lower computer and the safety skin through the TCP communication protocol, which is used for the safety skin to transmit contact trigger signals and distances to the lower computer.

[0110] S42. Turn on the external automatic operation of the industrial robot, execute the assembly task, and record the motion trajectory;

[0111] In this embodiment, step S42 is specifically as follows:

[0112] The operator sends an external automatic operation instruction through the upper computer UI interface in the industrial control computer to turn on the external automatic operation of the robot; after the external automatic operation is turned on, the industrial robot continuously transmits real-time positions to the upper computer, and the upper computer continuously stores the real-time trajectories of the robot within a fixed time window; and sends target positions to the robot at fixed time steps through the Kuka industrial robot RSI plugin package to control the robot to execute the assembly task and record the motion trajectory of the industrial robot.

[0113] S43. Activate the contact distance sensors arranged on the inner surface of the safety skin; predict the curved surface deformation of the safety skin at each time step during the external automatic operation of the industrial robot, and formulate a hierarchical control strategy for the robot on the premise of ensuring the safety of the operator until the assembly task is completed;

[0114] As a preferred implementation manner, step S43 in this embodiment specifically includes:

[0115] S431. After activating the contact distance sensors, the robot returns to the manually set safe Home point and determines that the safe Home point is the zero position where the safety skin is not affected by external forces; the control system performs initial calibration on each contact distance sensor and calibrates the contact distance shown by each sensor to L0; as Figure 4 shown, Figure 4 (a), (b), and (c) in

[0116] S432. Determine the surface deformation of the safety skin based on the real-time reading of the limited contact distance sensor and the estimated pose at the predicted safety skin points obtained according to the Kriging response model. When the safety skin undergoes surface deformation and the contact distance is less than L1 (as long as it occurs at one position), trigger the robot to decelerate.

[0117] S433. During the deceleration process of the industrial robot, continuously perform pose estimation at the predicted points and pose measurement at the sensor. When the safety skin undergoes surface deformation and the contact distance is less than L2 (as long as it occurs at one position), trigger the robot to stop immediately, and at the same time send a warning signal to the operator and the plant safety personnel.

[0118] S434. After triggering the robot to stop immediately, continuously judge the contact distance. If the contact distance at all positions has returned to L0, it is considered that the human-machine separation has occurred. After the operator confirms the human-machine separation in the upper computer interface, restart the external automatic operation of the robot. If it has not returned to L0 within the set emergency stop time threshold, guide the robot to retreat through the trajectory within the fixed time window stored in the upper computer.

[0119] In this embodiment, L0 = 350mm, L1 = 300mm, and L2 = 200mm are taken respectively; the emergency stop time threshold is set to 2s, and the fixed time window is set to 20s.

[0120] S435. Continuously judge the contact distance throughout the assembly task, and repeat steps S432 - S434 until the entire human-machine collaborative assembly task is finally completed.

[0121] In summary, through the real-time measurement and accurate prediction of the surface deformation of the safety skin, combined with the hierarchical control strategy safety protection mechanism, the present invention can ensure the interaction safety when the human and the machine collaborate to perform the assembly task, thereby improving the human-machine coexistence safety level and collaboration efficiency in the aerospace large-component assembly scenario.

[0122] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0123] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definable list of executable instructions for implementing a logical function, and can be embodied specifically in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in conjunction with these instruction execution systems, apparatuses, or devices.

[0124] The above embodiments have introduced the present invention in detail. Specific examples are used herein to elaborate on the principles and embodiments of the present invention. The description of the above embodiments is only for helping to understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A human-machine collaborative safety assembly method based on the safety skin of an industrial robot, characterized in that, Specifically, it includes the following steps: S1. Build a robotic industrial environment for human-robot collaborative assembly tasks; the robotic working environment includes: an industrial robot, a safety skin, an industrial control computer, and an operator; S2. Measure the actual poses of the industrial robot and the safety skin in real time, and construct a Kriging response model; S3. Solve the Kriging response model to predict the curved surface shape of the entire safety skin; S4. Design a safety protection strategy for the safety skin, select a corresponding robot control strategy according to the obtained curved surface shape, control the motion trajectory of the robot, and achieve the safe execution of the human-robot collaborative assembly task.

2. The human-machine collaborative safety assembly method based on the safety skin of an industrial robot according to claim 1, wherein, In step S1: The industrial robot is a heavy-duty industrial robot, which carries an end effector to receive external instructions and execute human-robot collaborative assembly tasks; The safety skin is wrapped around the outside of the industrial robot, and a limited number of contact distance sensors are randomly and dispersedly arranged on its inner surface to measure the pose data and force deformation of the safety skin; the number of positions where sensors are arranged is the same as the number of positions where sensors are not arranged; The operator is always arranged within the working range of the industrial robot, that is, it is set that there is a collision risk between the operator and the industrial robot; The industrial control computer is connected to the industrial robot and the safety skin through a network cable, and there is a communication connection between them. The industrial robot moves according to the point position information sent by the industrial control computer, and the industrial control computer receives the pose data sent by the safety skin and the industrial robot in real time to perceive the overall position of the system.

3. The human-machine collaborative safety assembly method based on the safety skin of an industrial robot according to claim 2, wherein, Step S2 specifically includes: S21. Take the base coordinate system of the industrial robot as the world coordinate system, and record the initial theoretical poses of each contact distance sensor and the positions where contact distance sensors are not arranged on the inner surface of the safety skin in the world coordinate system according to the robot's theoretical digital model and the actual installation position of the safety skin; The initial theoretical pose of the i-th contact distance sensor is expressed as: [x i , y i , z i , a i , b i , c i , i = 1, 2, ···, N; where x i , y i , z i are the initial three-dimensional coordinates of the i-th contact distance sensor, and a i , b i , c i are the initial three-axis rotation angles of the i-th contact distance sensor, and N is the total number of contact distance sensors; The initial theoretical pose of the t-th position where a contact distance sensor is not arranged is expressed as: wherein, is the initial three-dimensional coordinate at the t-th location where the contact distance sensor is not installed, is the initial three-axis rotation angle at the t-th location where the contact distance sensor is not installed; S22. During the operation of the safety skin, if the safety skin is deformed by external force, the contact distance sensors arranged on its inner surface are activated to obtain the actual measured poses at the positions of each sensor; then calculate the deviation between the actual measured pose and the initial theoretical pose, and the formula is expressed as: Among them, are respectively the actual measured pose, the initial theoretical pose, and the deviation between the actual measured pose and the initial theoretical pose of the contact distance sensor i; S23. Take the positions where contact distance sensors are not arranged on the inner surface of the safety skin as prediction points, and construct a Kriging response model to predict the curved surface deformation state of the entire safety skin; the benchmark input of the Kriging response model is the deviation between the actual measured pose of the contact distance sensor and the initial theoretical pose, and the output response value is used as the pose estimation deviation at the prediction point; the Kriging response model is expressed as a combination of a regression function and a stochastic process function, and the formula is expressed as: where f i (t) is the specified regression function, β i is the coefficient of the regression function, and z(t) is a stochastic process function; is the response value, i.e., the estimated pose deviation; S24. During the estimation process, to ensure the credibility of the estimation, the covariance is introduced as the evaluation criterion for the pose estimation error, and the stochastic process function is set to be a normal distribution function with a mean of 0 and a variance of σ 2 . Then the covariance between adjacent predicted points t1 and t2 is expressed as: Cov[z(t1),z(t2)] = σ 2 R(θ R ,t1,t2); where z(t1) and z(t2) are the stochastic process functions at the predicted points t1 and t2; R(θ R , t1, t2) is the Gaussian correlation function between two adjacent predicted points t1 and t2; the covariance between adjacent predicted points is used as the evaluation criterion for prediction accuracy, and the smaller the covariance, the higher the estimation accuracy of the model; the Gaussian correlation function is specifically expressed as: Among them, for two adjacent predicted points t1 and t2, and are respectively the components of the estimated pose deviation at t1 and t2 in the k-th degree of freedom.

4. A human-machine collaborative safety assembly method based on an industrial robot safety skin according to claim 3, characterized in that, Step S3 specifically includes: S31. Solve the Kriging model to obtain the estimated pose deviation of the prediction point as: where ω T is the weighting coefficient; ε is the measurement error of the Kriging response model; ζ is the matrix composed of the pose deviations of each contact distance sensor; the weighting coefficient is obtained by constructing and solving the Lagrangian function; S32. Denote the estimated pose deviation at any prediction point t obtained by solving the Kriging model as: Based on the estimated deviation of the prediction point and the initial theoretical pose, the estimated pose representation at the safe skin prediction point t in the world coordinate system at the current moment is calculated. That is, it satisfies: Among them, is the estimated three-dimensional coordinate at the prediction point t, is the estimated three-axis rotation angle at the prediction point t.

5. A human-machine collaborative safety assembly method based on an industrial robot safety skin according to claim 1, characterized in that, Step S4 specifically includes: S41. Establish communication between the industrial robot, the safety skin, and the industrial control computer; S42. Turn on the external automatic operation of the industrial robot, execute the assembly task and record the motion trajectory; S43. Activate the contact distance sensors arranged on the inner surface of the safety skin; predict the surface deformation of the safety skin at each time step during the external automatic operation of the industrial robot, and formulate a hierarchical control strategy for the robot on the premise of ensuring the safety of the operator until the assembly task is completed.

6. The human-machine collaborative safety assembly method based on the safety skin of an industrial robot according to claim 5, characterized in that, Specifically, step S41 is as follows: Establish communication between the upper and lower computers in the industrial control computer through the ADS communication protocol; establish communication between the lower computer and the industrial robot through the TCP communication protocol, and the upper computer sends instructions to the robot via the lower computer; establish communication between the upper computer and the industrial robot through the UDP communication protocol and the Kuka industrial robot RSI plugin package for the robot to transmit the real-time position to the upper computer and the upper computer to send the motion trajectory to the robot; establish communication between the lower computer and the safety skin through the TCP communication protocol for the safety skin to transmit the contact trigger signal and distance to the lower computer.

7. A human-robot collaborative safety assembly method based on the safety skin of an industrial robot according to claim 5, characterized in that Specifically, step S42 is as follows: The operator sends an external automatic operation instruction through the upper computer UI interface in the industrial control computer to start the external automatic operation of the robot; after the external automatic operation is started, the industrial robot continuously transmits the real-time position to the upper computer, and the upper computer continuously stores the real-time trajectory of the robot within a fixed time window; and sends the target position to the robot at a fixed time step interval through the Kuka industrial robot RSI plugin package to control the robot to perform the assembly task and record the motion trajectory of the industrial robot.

8. A human-machine collaborative safety assembly method based on an industrial robot safety skin according to claim 5, characterized in that Specifically, step S43 includes: S431. After activating the contact distance sensors, the robot returns to the manually set safe Home point and determines that the zero position where the safety skin is not affected by external forces at the safe Home point; the control system performs initial calibration on each contact distance sensor and calibrates the contact distance shown by each sensor to L0. S432. Determine the surface deformation of the safety skin according to the real-time readings of the limited contact distance sensors and the estimated pose at the predicted points of the safety skin obtained according to the Kriging response model; when the surface of the safety skin deforms and the contact distance is less than L1, trigger the robot to decelerate. S433. During the deceleration process of the industrial robot, continuously perform pose estimation at the predicted points and pose measurement at the sensors. When the surface of the safety skin deforms and the contact distance is less than L2, trigger the robot to stop suddenly and send a warning signal to the operator and the plant safety personnel at the same time. S434. After triggering the robot to stop suddenly, continuously judge the contact distance. If the contact distance at all positions has returned to L0, it is considered that the human and the machine are separated. After the operator confirms the separation of the human and the machine in the upper computer interface, restart the external automatic operation of the robot; if it has not returned to L0 within the set sudden stop time threshold, guide the robot to retreat through the trajectory stored in the upper computer within a fixed time window. S435. Keep judging the contact distance at all times during the entire assembly task, and continuously loop through steps S432 - S434 until the entire human-machine collaborative assembly task is finally completed.

Images