A robot soft force control interaction method, system, device and storage medium for a collaborative sewing scene

Abstract

The application discloses a kind of robot compliance force control interaction methods, systems, equipment and storage medium for collaborative sewing scene, comprising: preset desired contact force trajectory, velocity trajectory and pose trajectory;Before task starts, multiple groups of force sensor data are collected and dynamic gravity compensation is completed, real-time calculation real contact force;During task, real contact force is continuously acquired, velocity deviation and pose deviation are obtained, and real contact force, velocity deviation and pose deviation are input into compliance controller based on admittance model, dynamically adjust the target pose of next mechanical arm;Repeat execution compliance cooperation until the whole collaborative sewing task is completed;After task ends, parameter optimization algorithm based on smooth iterative learning is constructed, real contact force trajectory, velocity trajectory and pose trajectory are used as input, and compliance controller parameters are autonomously optimized by smooth iterative learning algorithm, to realize the best force-position trajectory tracking performance.The application realizes safe and stable robot collaborative sewing task.

Description

Technical Field

[0001] This invention belongs to the field of robot intelligent control technology, and in particular relates to a robot compliant force control interaction method, system, device and storage medium for autonomous sewing of flexible fabrics. Background Technology

[0002] The intelligent transformation of the textile manufacturing industry is accelerating the application of robotics in sewing processes. However, the traditional industrial robot operation mode based on rigid position control faces significant challenges in dealing with the special scenario of fabric sewing: the inherent flexibility and nonlinear deformation characteristics of fabric make the force-deformation coupling effect a key factor affecting sewing quality, often leading to process defects such as uneven stitches and fabric wrinkling. Especially in collaborative sewing scenarios, the dynamic interaction between the robot and the environment places higher demands on real-time force control accuracy and compliance.

[0003] Current technology faces two major bottlenecks: first, insufficient trajectory adaptability, with rigid trajectories struggling to cope with nonlinear fabric deformation, easily leading to tracking deviations or fabric damage; second, low security in collaborative interaction, lacking efficient force / position hybrid control strategies and unable to respond in real time to sudden changes in contact force. Existing impedance control or PID force control methods suffer from response lag or overshoot in complex scenarios, and fixed control parameters are difficult to adapt to diverse sewing task requirements. There is an urgent need to develop high-precision, adaptive, compliant force control methods to overcome the key technological bottlenecks in collaborative sewing. Summary of the Invention

[0004] Purpose of the invention: The purpose of this invention is to provide a robot compliant force control interaction method for collaborative sewing scenarios, so as to achieve good trajectory tracking performance and collaborative sewing effect, and have strong environmental adaptability.

[0005] Technical solution: The method described in this invention includes the following steps:

[0006] The preset contact force trajectory, robot velocity trajectory, and pose trajectory;

[0007] Before the mission begins, multiple sets of force sensor data are collected and dynamic gravity compensation is completed to calculate the real contact force between the end effector and the sewing platform in real time.

[0008] During the task, the actual contact force is continuously acquired. The actual robot velocity and pose trajectories are subtracted from the expected robot velocity and pose trajectories to obtain the velocity deviation and pose deviation. Then, the actual contact force, velocity deviation, and pose deviation are input into the compliant controller based on the admittance model to dynamically adjust the target pose of the robotic arm for the next step, thereby achieving compliant robot cooperation. The compliant cooperation is repeated until the entire collaborative sewing task is completed.

[0009] After the task is completed, a parameter optimization algorithm based on smooth iterative learning is constructed. The recorded real contact force trajectory, robot velocity trajectory and pose trajectory are used as input. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve the best force and position trajectory tracking performance.

[0010] Furthermore, multiple sets of force sensor data are collected and dynamic gravity compensation is performed to calculate the real contact force f(t) between the end effector and the sewing platform in real time, specifically including:

[0011] The robotic arm is placed in a free, static position, i.e., without external force contact. Then, the force sensor collects and records raw data f from N different static poses. s (t), N≥3; Measure the gravity value G of the end effector, and calculate the gravity component f in the force sensor coordinate system based on the pose information of the end effector. g (t); Calculate the zero-point data f0 of the force sensor using the least squares method; Calculate the actual contact force f(t) = f s (t)-f0-f g (t).

[0012] Furthermore, the actual contact force, velocity deviation, and pose deviation are input into a compliant controller based on an admittance model to dynamically adjust the target pose of the robotic arm in the next step; this includes the following steps:

[0013] (1) Constructing a parameterized robot compliant controller:

[0014]

[0015] in, This represents the increment of robot acceleration output by the compliant controller. and x e (t) represent the robot's velocity deviation and pose deviation, respectively, f d f(t) and f(t) represent the desired contact force and the actual contact force, respectively. m(t), b(t), and k(t) are parameters that change with time, representing the virtual inertia, damping, and stiffness parameters of the desired dynamic interaction process, respectively.

[0016] (2) The compliant controller updates the robot's target pose x(t+1) in real time:

[0017] Based on the acceleration increment output by the compliant controller, the adjustment amounts for velocity and pose are obtained through integration, thereby calculating the next robot target pose x(t+1), expressed by the formula:

[0018]

[0019] x t(t+1)=x(t+1)-x e (t+1)

[0020] in, and x e (t+1) represent the robot's velocity increment and pose increment at time t+1, respectively, and x(t+1) and x t (t+1) represents the robot's true pose and target pose at time t+1, respectively, and △t is the sampling time.

[0021] Furthermore, a parameter optimization algorithm based on smooth iterative learning is constructed. Taking the recorded real contact force trajectory, robot velocity trajectory, and pose trajectory as input, the smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve optimal force-position trajectory tracking performance. This includes the following steps:

[0022] (1) Construct the state-space equation of the parameterized compliant controller and convert it into a discrete expression form to aid in the derivation of the iterative learning controller;

[0023] (2) Calculate the controller output at the next moment based on the discrete expression of the robot compliant controller, and derive the controller error correction expression;

[0024] (3) Design an adaptive iterative learning update law and construct a data-driven environmental stiffness estimation algorithm;

[0025] (4) Design a smooth update method based on the principle of bilinear interpolation to enhance the continuity of the iterative process, and combine it with the adaptive iterative learning update law to form a parameter optimization algorithm based on smooth iterative learning.

[0026] Furthermore, step (1) specifically includes the following steps:

[0027] (11) Let the environment model of the interaction between the robot's end effector and the sewing platform be f(t) = k c (t)x(t), establish the state-space equation of the control system:

[0028]

[0029] Where, f(t), f d (t) represents the trajectories of the actual contact force and the expected contact force, respectively, and k c x(t) represents the environmental stiffness, and x(t), These are the actual robot pose trajectory and velocity trajectory, respectively. and x e (t+1) represents the robot's velocity increment and pose increment at time t+1, respectively; m(t), b(t), and k(t) are time-varying parameters, representing the virtual inertia, damping, and stiffness parameters of the desired dynamic interaction process, respectively; xd (t), These represent the desired robot pose trajectory and velocity trajectory, respectively, x e (t), These represent pose deviation and velocity deviation, respectively, with Δt being the sampling time.

[0030] (12) Transform the state-space equation of the robot compliant controller into a typical discrete expression:

[0031] z(t+1)=(βu T (t)A+B)(r(t)-y(t))

[0032] y(t)=C(t)z(t)+D(t)r(t)

[0033] Where z(t+1) is the state vector composed of velocity deviation and pose deviation at time t+1, β, A, and B are all intermediate variables, and u T (t) is the transpose of the vector of controller parameters at time t, C(t) is the matrix of environmental stiffness parameters at time t, D(t) is the matrix of intermediate variables at time t, z(t) is the state vector composed of velocity deviation and pose deviation at time t, y(t) is the output vector composed of target velocity, pose and contact force at time t, and r(t) is the input vector composed of desired contact force and robot motion trajectory at time t.

[0034] Furthermore, step (2) specifically includes the following steps:

[0035] (21) Let the compliant controller parameter matrix U(t) = βu T (t)A+E, where β, A, and E are all intermediate variables, according to the discrete expression of the robot compliant controller, the output vector expression at time t+1 is:

[0036] y(t+1)=U(t)(r(t)-y(t))+D(t+1)r(t+1)

[0037] Where y(t+1) is the output vector composed of the target velocity, pose and contact force at time t+1, D(t+1) is the intermediate variable matrix at time t+1, r(t) is the input vector composed of the expected contact force and robot trajectory at time t, y(t) is the output vector composed of the target velocity, pose and contact force at time t, and r(t+1) is the vector composed of the expected contact force and robot trajectory at time t+1.

[0038] (22) Construct an error correction term ΔU(t). After adding the error correction term, the system output at time t+1 is compensated to the reference value. Construct the controller output error vector e(t+1) = r(t+1) - y(t+1) at time t+1. Then, according to the system output vector expression and the error correction term, we get:

[0039] y(t+1)+e(t+1)=C(t+1)(U(t)+△U(t))e(t)+D(t+1)r(t+1)

[0040] Where C(t+1) is the environmental stiffness parameter matrix at t+1, and e(t) is the output vector error of the controller at time t;

[0041] Subtracting the two formulas above and using the concept of the generalized inverse of a matrix, we obtain the error correction expression:

[0042] △U(t)=C + (t+1)e(t+1)e + (t)

[0043] Among them, C + (t+1) is the generalized inverse of the environmental stiffness parameter matrix at t+1, e + (t) is the generalized inverse of the controller's output error vector at time t.

[0044] Furthermore, step (3) specifically includes the following steps:

[0045] (31) Based on the characteristic of iterative learning algorithms continuously approaching the convergence region, the parameter update law of the robot compliant controller based on iterative learning can be constructed by multiplying the expression of △U(t) by the learning rate α:

[0046]

[0047] Among them, U k+1 (t) is the compliant controller parameter matrix for the (k+1)th iteration, U k (t) is the compliant controller parameter matrix for the k-th iteration, ΔU k (t) represents the error correction term for the k-th iteration. Let e ​​be the generalized inverse of the environmental stiffness parameter matrix in the k-th iteration. k (t+1) is the output error vector of the controller in the k-th iteration. It is the generalized inverse of the output error vector of the controller in the k-th iteration;

[0048] (32) Construct an adaptive learning rate α, whose adaptive expression is:

[0049]

[0050] Where |e(t)| is the magnitude of the controller's output error vector at time t in the current iteration round;

[0051] (33) In the parameter update law of the robot compliant controller based on iterative learning, C(t) contains the unknown environmental stiffness parameter matrix. Construct the corresponding estimation matrix. With respect to the estimated cost function J(C(t)):

[0052]

[0053] Among them, y k-1 (t) is the output vector composed of the target velocity, pose, and contact force in the (k-1)th iteration, z k-1 (t) is the state vector composed of the velocity deviation and pose deviation in the (k-1)th iteration. This is the intermediate variable matrix, where μ is the weighting factor. Let r(t) be the estimated environmental stiffness parameter matrix for the (k-1)th iteration, and r(t) be the input vector consisting of the expected contact force and the robot's motion trajectory at time t.

[0054] make The estimated environmental stiffness parameters are calculated, and the expression is:

[0055]

[0056] Among them, T D η is the intermediate variable matrix, where 0 < η ≤ 2 is the step size factor, and μ is the weight factor.

[0057] Furthermore, step (4) specifically includes the following steps:

[0058] (41) Assume that the output of the control system at time t is y(t), the output of the control system closest to it on the left is y(t-1), and the output of the control system closest to it on the right is y(t+1). Obviously, there are error correction terms ΔU at times t-1 and t+1 in the k-th iteration. k (t-1) and △U k (t+1), then a smooth update method for the error correction term is designed based on the bilinear interpolation principle, and the expression is:

[0059]

[0060] Among them, U k+1 (t) is the compliant controller parameter matrix for the (k+1)th iteration, U k (t) is the compliant controller parameter matrix for the k-th iteration;

[0061] (42) Similar to the smooth update process of the error correction term, the smooth update expression for the environmental stiffness parameter is:

[0062]

[0063] Among them, C k+1 (t) is the environmental stiffness parameter matrix for the (k+1)th iteration, C k (t) is the environmental stiffness parameter matrix for the k-th iteration, ΔC k (t+1) represents the deviation of the environmental stiffness parameter matrix between the k-th iteration and the (k-1)-th iteration at time t+1, ΔC k (t-1) represents the deviation of the environmental stiffness parameter matrix between the k-th iteration and the (k-1)-th iteration at time t-1;

[0064] (43) The complete parameter optimization algorithm based on smooth iterative learning is as follows:

[0065] z k (t+1)=U k (t)(r(t)-y k (t))

[0066] y k (t)=C(t)z k (t)+D(t)r(t)

[0067]

[0068] Among them, z k (t+1) is the state vector composed of velocity deviation and pose deviation at time t+1 of the k-th iteration, z k (t) is the state vector composed of the velocity deviation and pose deviation at time t of the k-th iteration. k (t) is the output vector composed of the target velocity, pose, and contact force at time t of the k-th iteration. k-1 (t) represents the compliant controller parameter matrix at time t in the (k-1)th iteration, where α is the adaptive learning rate. Let e ​​be the generalized inverse of the environmental stiffness parameter matrix at time t+1 of the (k-1)th iteration. k-1 (t+1) is the output error vector of the controller in the (k-1)th iteration. Let $\mathbf{k-1}$ be the generalized inverse of the output error vector of the controller in the (k-1)th iteration. Let y be the estimated environmental stiffness parameter matrix for the k-th iteration. k-1 (t) is the output vector composed of the target velocity, pose, and contact force in the (k-1)th iteration, z k-1 (t) is the state vector composed of the velocity deviation and pose deviation in the (k-1)th iteration. Let T be the estimated environmental stiffness parameter matrix for the (k-1)th iteration. D Let be the intermediate variable matrix, 0 < η ≤ 2 be the step size factor, and μ be the weight factor.

[0069] The system of the present invention includes:

[0070] The force-position operation trajectory preset unit is used to preset the contact force trajectory, robot velocity trajectory, and pose trajectory.

[0071] The real contact force calculation unit is used to collect multiple sets of force sensor data and complete dynamic gravity compensation before the task begins, and to calculate the real contact force between the end effector and the sewing platform in real time.

[0072] The collaborative sewing unit continuously acquires real contact force during the task. It subtracts the actual robot velocity and pose trajectories from the expected robot velocity and pose trajectories to obtain velocity and pose deviations. Then, it inputs the real contact force, velocity deviation, and pose deviation into a compliant controller based on an admittance model to dynamically adjust the target pose of the robotic arm for the next step, thereby achieving compliant robot collaboration. The compliant collaboration is repeated until the entire collaborative sewing task is completed.

[0073] The parameter optimization unit is used to construct a parameter optimization algorithm based on smooth iterative learning after the task is completed. The recorded real contact force trajectory, robot velocity trajectory and pose trajectory are used as inputs. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve the best force and position trajectory tracking performance.

[0074] The electronic device of the present invention includes a memory, a processor, and a computer program / instruction stored in the memory and executable on the processor. When the computer program / instruction is executed by the processor, it implements the steps of the robot compliant force control interaction method for collaborative sewing scenarios.

[0075] Beneficial effects: Compared with the prior art, the significant technical effects of the present invention are: it can provide a complete set of effective and universal robot compliant force control interaction algorithms, effectively solve the performance problems of force trajectory tracking and the safety interaction between the robot and the environment during collaborative sewing, realize the autonomous optimization of robot compliant control parameters during sewing, improve the tracking performance and actual sewing effect of the collaborative process, and can be widely applied to various different production and manufacturing processes, with good practicality and application prospects. Attached Figure Description

[0076] Figure 1 This is a schematic diagram of the robot compliant force control interaction process of the present invention;

[0077] Figure 2 This is a diagram of the variable parameter compliant control structure based on iterative learning according to the present invention. Detailed Implementation

[0078] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

[0079] This invention proposes a high-precision, adaptive robot compliant force control interaction method for collaborative sewing scenarios, aiming to solve the key technical challenges of force-position coordinated control in collaborative sewing and providing reliable technical support for the intelligent upgrading of the textile industry. This method abstracts the robot collaborative sewing process into a problem of tracking the contact force trajectory and the robot's motion trajectory. It reconstructs the compliant controller through a parameterization method and proposes a parameter optimization method based on iterative learning to obtain the optimal control parameters during the force / position operation process, thereby achieving good trajectory tracking performance and collaborative sewing effect, and possessing strong environmental adaptability.

[0080] like Figure 1 As shown, the specific steps of the method described in this invention are as follows:

[0081] S1. Preset the desired contact force trajectory and robot motion trajectory, including the desired contact force f between the end effector and the sewing platform. d (t); the desired speed of the robot in a given sewing task. and pose trajectory x d (t).

[0082] S2. Before the task begins, collect multiple sets of force sensor data and complete dynamic gravity compensation. Calculate the real contact force f(t) between the end effector and the sewing platform in real time, specifically including:

[0083] A force sensor is installed between the robotic arm flange and the end effector. The robotic arm is placed in a free state (i.e., no external force contact) at rest. Then, raw data f from the force sensor is collected and recorded at N (N≥3) different static poses. s (t); Measure the gravity value G of the end effector, and calculate the gravity component f in the force sensor coordinate system based on the pose information of the end effector. g (t); Calculate the zero-point data f0 of the force sensor using the least squares method; Calculate the actual contact force f(t) = f s (t)-f0-f g (t).

[0084] S3. During the task, continuously acquire the actual contact force f(t) and utilize the desired robot velocity trajectory. and pose trajectory x d (t) minus the actual robot velocity trajectory The velocity deviation is obtained from the pose trajectory x(t). and pose deviation x e(t)=x d (t)-x(t), then the actual contact force f(t) and velocity deviation are calculated. and pose deviation x e (t) Input the compliant controller based on the admittance model to dynamically adjust the next target pose of the robotic arm x(t+1), thereby achieving compliant cooperation of the robot.

[0085] Further, step S3 includes the following steps:

[0086] S31. Construct a parameterized robot compliance controller:

[0087]

[0088] in, This represents the increment of robot acceleration output by the compliant controller. and x e (t) represent the robot's velocity deviation and pose deviation, respectively, f d f(t) and f(t) represent the desired contact force and the actual contact force, respectively. m(t), b(t), and k(t) are time-varying parameters, representing the virtual inertia, damping, and stiffness parameters of the desired dynamic interaction process, respectively.

[0089] S32, The compliant controller updates the robot's target pose x(t+1) in real time:

[0090] Based on the acceleration increment output by the compliant controller, the adjustment amounts for velocity and pose are obtained through integration, thereby calculating the next robot target pose x(t+1), expressed by the formula:

[0091]

[0092] x t (t+1)=x(t+1)-x e (t+1)

[0093] in, and x e (t+1) represent the robot's velocity increment and pose increment at time t+1, respectively, and x(t+1) and x t (t+1) represents the robot's true pose and target pose at time t+1, respectively, and △t is the sampling time.

[0094] S4. Repeat the smooth collaborative operation of step S3 until the entire collaborative sewing task is completed.

[0095] S5. After the task is completed, a parameter optimization algorithm based on smooth iterative learning is constructed. The recorded real contact force trajectory, robot velocity trajectory, and pose trajectory are used as inputs. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve optimal force-position trajectory tracking performance. Specifically, this includes the following steps:

[0096] S51. Construct the state-space equation of the parameterized compliant controller and convert it into a discrete form to aid in the derivation of the iterative learning controller;

[0097] Further, step S51 includes the following steps:

[0098] S511. Let the environment model for the interaction between the robot's end effector and the sewing platform be f(t) = k. c The state-space equation of the control system can be established using x(t)(t):

[0099]

[0100] Where, k c (t) represents environmental stiffness;

[0101] S512. Convert the state-space equation of the robot compliant controller into a typical discrete expression:

[0102] z(t+1)=(βu T (t)A+B)(r(t)-y(t))

[0103] y(t)=C(t)z(t)+D(t)r(t)

[0104] in, z(t+1) is the state vector composed of velocity deviation and pose deviation at time t+1, β, A, and B are intermediate variables, D(t) is the intermediate variable matrix at time t, and u T C(t) is the transpose of the vector of controller parameters at time t, and C(t) is the matrix of environmental stiffness parameters at time t.

[0105] S52. Calculate the controller output at the next moment based on the discrete expression of the robot compliant controller, and derive the controller error correction expression.

[0106] Further, step S52 includes the following steps:

[0107] S521, Let the compliant controller parameter matrix U(t) = βu T (t)A+E, based on the discrete expression of the robot compliant controller, the output vector expression at time t+1 can be obtained as follows:

[0108] y(t+1)=U(t)(r(t)-y(t))+D(t+1)r(t+1)

[0109] Where y(t+1) is the output vector composed of the target velocity, pose and contact force at time t+1, D(t+1) is the intermediate variable matrix at time t+1, and r(t+1) is the vector composed of the expected contact force and robot trajectory at time t+1.

[0110] S522. Construct an error correction term ΔU(t). After adding the error correction term, the system output at time t+1 is compensated to the reference value. Construct the controller output error vector e(t+1) = r(t+1) - y(t+1) at time t+1. Then, according to the system output vector expression and the error correction term, we can obtain:

[0111] y(t+1)+e(t+1)=C(t+1)(U(t)+△U(t))e(t)+D(t+1)r(t+1)

[0112] Where C(t+1) is the environmental stiffness parameter matrix at time t+1, and e(t) is the output vector error of the controller at time t. Subtracting the two formulas above and using the concept of matrix generalized inverse, we obtain the error correction expression:

[0113] △U(t)=C + (t+1)e(t+1)e + (t)

[0114] Among them, C + (t+1) is the generalized inverse of the environmental stiffness parameter matrix at t+1, e + (t) is the generalized inverse of the controller's output error vector at time t.

[0115] S53. Design an adaptive iterative learning update law and construct a data-driven environmental stiffness estimation algorithm.

[0116] Figure 2 This is a diagram of the variable parameter compliant control structure based on iterative learning according to the present invention.

[0117] Further, step S53 includes the following steps:

[0118] S531. Based on the characteristic of iterative learning algorithms continuously approaching the convergence region, the parameter update law of a robot compliant controller based on iterative learning can be constructed by multiplying the expression of ΔU(t) by the learning rate α:

[0119]

[0120] Among them, U k+1 (t) is the compliant controller parameter matrix for the (k+1)th iteration, Uk (t) is the compliant controller parameter matrix for the k-th iteration, ΔU k (t) represents the error correction term for the k-th iteration. Let e ​​be the generalized inverse of the environmental stiffness parameter matrix in the k-th iteration. k (t+1) is the output error vector of the controller in the k-th iteration. It is the generalized inverse of the output error vector of the controller in the k-th iteration.

[0121] S532. Construct an adaptive learning rate α to improve the convergence speed of the iterative learning process. The adaptive learning rate α is highly correlated with the output error vector of the control system, and its adaptive expression is:

[0122]

[0123] Where |e(t)| is the magnitude of the controller's output error vector at time t in the current iteration round.

[0124] S533. In the parameter update law of a robot compliant controller based on iterative learning, C(t) contains an unknown environmental stiffness parameter matrix. Construct the corresponding estimation matrix. With respect to the estimated cost function J(C(t)):

[0125]

[0126] Among them, y k-1 (t) is the output vector composed of the target velocity, pose, and contact force in the (k-1)th iteration, z k-1 (t) is the state vector composed of the velocity deviation and pose deviation in the (k-1)th iteration. This is the intermediate variable matrix, where μ is the weighting factor. This is the estimated environmental stiffness parameter matrix for the (k-1)th iteration.

[0127] make The estimated environmental stiffness parameters can be calculated, and the expression is:

[0128]

[0129] Among them, T D η is the intermediate variable matrix, where 0 < η ≤ 2 is the step size factor, and μ is the weight factor.

[0130] S54. Design a smooth update method based on the principle of bilinear interpolation to enhance the continuity of the iterative process, and combine it with the adaptive iterative learning update law to form a parameter optimization algorithm based on smooth iterative learning.

[0131] S541. Assume the output of the control system at time t is y(t), the output of the control system closest to it on the left is y(t-1), and the output of the control system closest to it on the right is y(t+1). Obviously, there are error correction terms ΔU at times t-1 and t+1 in the k-th iteration. k (t-1) and △U k (t+1), then based on the principle of bilinear interpolation, a smooth update method for the error correction term can be designed, with the expression:

[0132]

[0133] S542. Similar to the smooth update process of the error correction term, the smooth update expression for the environmental stiffness parameter is:

[0134]

[0135] Among them, C k+1 (t) is the environmental stiffness parameter matrix for the (k+1)th iteration, C k (t) is the environmental stiffness parameter matrix for the k-th iteration, ΔC k (t+1) represents the deviation of the environmental stiffness parameter matrix between the k-th iteration and the (k-1)-th iteration at time t+1, ΔC k (t-1) represents the deviation of the environmental stiffness parameter matrix between the k-th iteration and the (k-1)-th iteration at time t-1;

[0136] S543. The complete parameter optimization algorithm based on smooth iterative learning is as follows:

[0137] z k (t+1)=U k (t)(r(t)-y k (t))

[0138] y k (t)=C(t)z k (t)+D(t)r(t)

[0139]

[0140] Among them, z k (t+1) is the state vector composed of velocity deviation and pose deviation at time t+1 of the k-th iteration, z k (t) is the state vector composed of the velocity deviation and pose deviation at time t of the k-th iteration. k (t) is the output vector composed of the target velocity, pose, and contact force at time t of the k-th iteration. k-1 (t) represents the compliant controller parameter matrix at time t in the (k-1)th iteration, where α is the adaptive learning rate. Let e ​​be the generalized inverse of the environmental stiffness parameter matrix at time t+1 of the (k-1)th iteration. k-1 (t+1) is the output error vector of the controller in the (k-1)th iteration. It is the generalized inverse of the output error vector of the controller in the (k-1)th iteration.

[0141] The method of this invention achieves safe and stable robotic collaborative sewing tasks through precise force and position control, providing a key fundamental solution for robotic autonomous sewing technology, and has significant economic value and application prospects for promoting the intelligent upgrading of the textile industry.

[0142] The system of the present invention includes:

[0143] The force-position operation trajectory preset unit is used to preset the contact force trajectory, robot velocity trajectory, and pose trajectory.

[0144] The real contact force calculation unit is used to collect multiple sets of force sensor data and complete dynamic gravity compensation before the task begins, and to calculate the real contact force between the end effector and the sewing platform in real time.

[0145] The collaborative sewing unit continuously acquires real contact force during the task. It subtracts the actual robot velocity and pose trajectories from the expected robot velocity and pose trajectories to obtain velocity and pose deviations. Then, it inputs the real contact force, velocity deviation, and pose deviation into a compliant controller based on an admittance model to dynamically adjust the target pose of the robotic arm for the next step, thereby achieving compliant robot collaboration. The compliant collaboration is repeated until the entire collaborative sewing task is completed.

[0146] The parameter optimization unit is used to construct a parameter optimization algorithm based on smooth iterative learning after the task is completed. The recorded real contact force trajectory, robot velocity trajectory and pose trajectory are used as inputs. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve the best force and position trajectory tracking performance.

[0147] The electronic device of the present invention includes a memory, a processor, and a computer program / instruction stored in the memory and executable on the processor. When the computer program / instruction is executed by the processor, it implements the steps of the robot compliant force control interaction method for collaborative sewing scenarios.

[0148] The computer-readable storage medium of the present invention stores computer instructions, which, when invoked, are used to execute the steps of the robot compliant force control interaction method for collaborative sewing scenarios.

[0149] The foregoing has described the relevant content of the present invention. Those skilled in the art will be able to implement the present invention based on these descriptions. All other embodiments obtained by those skilled in the art based on the above description of the present invention without inventive effort should fall within the scope of protection of the present invention.

Claims

1. A robot compliant force control interaction method for collaborative sewing scenarios, characterized in that, Includes the following steps: Preset the desired contact force trajectory, robot velocity trajectory, and pose trajectory; Before the mission begins, multiple sets of force sensor data are collected and dynamic gravity compensation is completed to calculate the real contact force between the end effector and the sewing platform in real time. During the task, the actual contact force is continuously acquired. The actual robot velocity and pose trajectories are subtracted from the expected robot velocity and pose trajectories to obtain the velocity deviation and pose deviation. Then, the actual contact force, velocity deviation, and pose deviation are input into the compliant controller based on the admittance model to dynamically adjust the target pose of the robotic arm for the next step, thereby achieving compliant robot cooperation. The compliant cooperation is repeated until the entire collaborative sewing task is completed. After the task is completed, a parameter optimization algorithm based on smooth iterative learning is constructed. The recorded real contact force trajectory, robot velocity trajectory, and pose trajectory are used as inputs. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve optimal force-position trajectory tracking performance, including the following steps: (1) Construct the state-space equation of the parameterized compliant controller and convert it into a discrete expression to aid in the derivation of the iterative learning controller; specifically including: (11) The environment model for the interaction between the robot's end effector and the sewing platform is as follows: Establish the state-space equations of the control system: ； in, , These are the trajectories of the actual contact force and the expected contact force, respectively. Indicates environmental stiffness. , These are the actual robot pose trajectory and velocity trajectory, respectively. and They represent The robot's velocity increment and pose increment at any given moment. , , These are all time-varying parameters, representing the virtual inertia, damping, and stiffness parameters of the desired dynamic interaction process, respectively. , These represent the desired robot pose trajectory and velocity trajectory, respectively. , These are pose deviation and velocity deviation, respectively. Sampling time; (12) Transform the state-space equation of the robot compliant controller into a typical discrete expression: ； in, for The state vector composed of the instantaneous velocity deviation and pose deviation. , and All are intermediate variables. Let be the transpose of the vector containing the controller parameters at time t. Let be the environmental stiffness parameter matrix at time t. Let be the intermediate variable matrix at time t. Let be the state vector composed of velocity deviation and pose deviation at time t. Let be the output vector composed of the target velocity, pose, and contact force at time t. Let t be the input vector consisting of the expected contact force and the robot's trajectory at time t; (2) Calculate the controller output at the next moment based on the discrete expression of the robot compliant controller, and derive the controller error correction expression; (3) Design an adaptive iterative learning update law and construct a data-driven environmental stiffness estimation algorithm; (4) Design a smooth update method based on the principle of bilinear interpolation to enhance the continuity of the iterative process, and combine it with the adaptive iterative learning update law to form a parameter optimization algorithm based on smooth iterative learning.

2. The robot compliant force control interaction method for collaborative sewing scenarios according to claim 1, characterized in that, Collect data from multiple force sensors and perform dynamic gravity compensation to calculate the actual contact force between the end effector and the sewing platform in real time. Specifically, it includes: The robotic arm is placed in a free, stationary position, i.e., without external force contact. Subsequently, the force sensor collects and records raw data from N different static poses. N≥3; Measure the gravity value of the end effector. The gravity component in the force sensor coordinate system is calculated based on the pose information of the end effector. ; Calculate the zero-point data of the force sensor using the least squares method ; Calculate the actual contact force .

3. The robot compliant force control interaction method for collaborative sewing scenarios according to claim 1, characterized in that, The actual contact force, velocity deviation, and pose deviation are input into a compliant controller based on an admittance model to dynamically adjust the target pose of the robotic arm for the next step; including the following steps: (1) Constructing a parameterized robot compliant controller: ； in, This represents the increment of robot acceleration output by the compliant controller. and These represent the robot's velocity deviation and pose deviation, respectively. and These represent the expected contact force and the actual contact force, respectively. , , These are all time-varying parameters, representing the virtual inertia, damping, and stiffness parameters of the desired dynamic interaction process, respectively. (2) The compliant controller updates the robot's target pose for the next step in real time. : Based on the acceleration increment output by the compliant controller, the adjustment amounts for velocity and pose are obtained through integration, thereby calculating the robot's target pose for the next step. The formula is expressed as: ； in, and They represent The robot's velocity increment and pose increment at any given moment. and They represent The robot's actual pose and the target pose at any given moment. Sampling time.

4. The robot compliant force control interaction method for collaborative sewing scenarios according to claim 1, characterized in that, Step (2) specifically includes the following steps: (21) Set the compliant controller parameter matrix , , and All are intermediate variables. Based on the discrete expression of the robot compliant controller, we obtain... The expression for the output vector at time step is: ； in, for The output vector is composed of the target velocity, pose, and contact force at any given moment. for Intermediate variable matrix at time step, Let be the input vector consisting of the expected contact force and the robot's trajectory at time t. Let be the output vector composed of the target velocity, pose, and contact force at time t. for A vector consisting of the expected contact force and the robot's trajectory at any given moment; (22) Construction error correction term After the error correction term is added, the system is able to... The output at each time step is compensated to the reference value; construction Output error vector of the time controller Then, based on the system output vector expression and the error correction term, we get: ； in, for The environmental stiffness parameter matrix at time t, Let be the output vector error of the controller at time t; Subtracting the two formulas above and using the concept of the generalized inverse of a matrix, we obtain the error correction expression: ； in, for The generalized inverse of the environmental stiffness parameter matrix at time t, Let be the generalized inverse of the controller's output error vector at time t.

5. The robot compliant force control interaction method for collaborative sewing scenarios according to claim 1, characterized in that, Step (3) specifically includes the following steps: (31) Based on the characteristic of iterative learning algorithms to continuously approach the convergence region, Multiplying the expression by the learning rate φ, we can construct the parameter update law for a robot compliant controller based on iterative learning: ； in, For the first The compliant controller parameter matrix for the next iteration For the first The compliant controller parameter matrix for the next iteration For the first Error correction term for the next iteration For the first The generalized inverse of the environmental stiffness parameter matrix of the next iteration For the first The output error vector of the controller in the next iteration. For the first The generalized inverse of the controller's output error vector in the next iteration; (32) Construct an adaptive learning rate φ, whose adaptive expression is: ； in, Let be the magnitude of the controller's output error vector at time t in the current iteration round; (33) In the parameter update law of the robot compliant controller based on iterative learning, 𝐶(𝑡) contains the unknown environmental stiffness parameter matrix, and the corresponding estimation matrix is ​​constructed. With the estimated cost function : ； in, For the first The output vector of the next iteration consists of the target velocity, pose, and contact force. For the first The state vector is composed of the velocity deviation and pose deviation of each iteration. For the intermediate variable matrix, As a weighting factor, For the first The estimated environmental stiffness parameter matrix of the next iteration. Let t be the input vector consisting of the expected contact force and the robot's trajectory at time t; make The estimated environmental stiffness parameters are calculated, and the expression is: ； in, For the intermediate variable matrix, Step size factor; This is the weighting factor.

6. The robot compliant force control interaction method for collaborative sewing scenarios according to claim 1, characterized in that, Step (4) specifically includes the following steps: (41) Assume that the output of the control system at time t is The output of the control system closest to it on the left is The output of the control system closest to the right is It is obvious that there exists in the k-th iteration. and Error correction term for time and Then, based on the principle of bilinear interpolation, a smooth update method for the error correction term is designed, and its expression is: ； in, For the first The compliant controller parameter matrix for the next iteration For the first The compliant controller parameter matrix for the next iteration; (42) Similar to the smooth update process of the error correction term, the smooth update expression for the environmental stiffness parameter is: ； in, For the first The environmental stiffness parameter matrix of the next iteration For the first The environmental stiffness parameter matrix of the next iteration For the first Next and first iteration Deviation of environmental stiffness parameter matrix at time t, For the first Next and first iteration Deviation of the environmental stiffness parameter matrix at any given time; (43) The complete parameter optimization algorithm based on smooth iterative learning is as follows: ； in, For the first iteration The state vector composed of the velocity deviation and pose deviation at any given time. For the first iteration The state vector composed of the velocity deviation and pose deviation at any given time. For the first The output vector at time t of the next iteration consists of the target velocity, pose, and contact force. For the first The compliant controller parameter matrix at time t of the next iteration For adaptive learning rate, For the first iteration The generalized inverse of the environmental stiffness parameter matrix at any given time. For the first The output error vector of the controller in the next iteration. For the first The generalized inverse of the output error vector of the next iteration controller For the first The estimated environmental stiffness parameter matrix of the next iteration. For the first The output vector of the next iteration consists of the target velocity, pose, and contact force. For the first The state vector is composed of the velocity deviation and pose deviation of each iteration. For the first The estimated environmental stiffness parameter matrix of the next iteration. For the intermediate variable matrix, Step size factor This is the weighting factor.

7. A system for a robot compliant force control interaction method for collaborative sewing scenarios as described in any one of claims 1-6, characterized in that, include: The force-position operation trajectory preset unit is used to preset the desired contact force trajectory, robot velocity trajectory, and pose trajectory; The real contact force calculation unit is used to collect multiple sets of force sensor data and complete dynamic gravity compensation before the task begins, and to calculate the real contact force between the end effector and the sewing platform in real time. The collaborative sewing unit continuously acquires real contact force during the task. It subtracts the actual robot velocity and pose trajectories from the expected robot velocity and pose trajectories to obtain velocity and pose deviations. Then, it inputs the real contact force, velocity deviation, and pose deviation into a compliant controller based on an admittance model to dynamically adjust the target pose of the robotic arm for the next step, thereby achieving compliant robot collaboration. The compliant collaboration is repeated until the entire collaborative sewing task is completed. The parameter optimization unit is used to construct a parameter optimization algorithm based on smooth iterative learning after the task is completed. The recorded real contact force trajectory, robot velocity trajectory and pose trajectory are used as inputs. The smooth iterative learning algorithm autonomously optimizes the compliant controller parameters to achieve the best force and position trajectory tracking performance.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program / instructions stored in the memory and executable on the processor, wherein the computer program / instructions, when executed by the processor, implement the steps of the robot compliant force control interaction method for collaborative sewing scenarios according to any one of claims 1-6.

Images