Alternating impact disassembly operation robot dynamic characteristic modeling method and system

Abstract

The application relates to an alternating impact disassembly operation robot dynamic characteristic modeling method and system, and relates to the technical field of robot control. Firstly, the dynamic characteristics of the robot joint side and the motor side are comprehensively considered, a flexible joint dynamics model of torque driving is established, and the influence of joint flexibility on impact response is determined. Then, an exponential inelastic continuous contact model estimation method is used to estimate the radial impact force and the tangential friction force, an alternating impact dynamics model is constructed, the momentum change and the nonlinear dynamics effect in the impact process are quantified. Finally, a multi-space dynamic characteristic coupling model is established, the interaction mechanism of the joint space, the impact contact space and the task space is systematically analyzed, the accurate post-impact speed is output, the impact force control strategy is generated, the dynamic response accuracy of the robot under the alternating impact operation is improved, the active utilization of the impact effect of the robot is realized, and the stability, accuracy and efficiency of the disassembly operation are ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, in particular to a dynamic characteristic modeling method and system of an alternating impact disassembly operation robot. BACKGROUND

[0002] At present, the disassembly of power battery packs is mainly manual violent disassembly, which not only has high labor intensity and high safety risk, but also may cause environmental pollution, which is not conducive to the sustainable development of the industry. Therefore, developing an intelligent and automatic power battery fine disassembly robot system has become an inevitable choice to improve the recovery efficiency and safety. The disassembly and recovery of retired power battery packs is a reverse process of the battery pack grouping process, involving multiple links such as unpacking, glue removal, cutting, and crushing.

[0003] In actual operation, on the one hand, due to the high disassembly beat requirement, the disassembly tool impacts the components at a high speed; on the other hand, the state of the retired battery pack varies greatly, and there are situations such as screw damage, metal component rust, and glue solidification, which are difficult to deal with efficiently by traditional mechanical disassembly methods, and usually alternating impact methods are used for disassembly, that is, high-frequency impact force is applied in a short time to achieve effective disassembly. However, the alternating impact process involves complex mechanical properties, especially when the robot autonomously performs a disassembly task, how to accurately model and control the impact force becomes a key problem affecting the stability and accuracy of the operation.

[0004] Therefore, the existing alternating impact disassembly operation robot lacks an accurate impact dynamics model, resulting in instability and low control accuracy of the impact force in the disassembly operation. SUMMARY

[0005] The present application provides a dynamic characteristic modeling method and system of an alternating impact disassembly operation robot, aiming to build an accurate impact dynamics model. The accurate impact force is obtained in real time during the robot performing a disassembly operation task, an impact force control strategy is introduced, the active use of the impact by the robot is realized, the stability and accuracy of the impact force control are improved, the stability, accuracy, and efficiency of the disassembly operation are ensured, and the problem of insufficient impact force control in the disassembly operation due to the lack of an accurate impact dynamics model in the prior art is solved.

[0006] In a first aspect, the present application provides a dynamic characteristic modeling method of an alternating impact disassembly operation robot, comprising:

[0007] The dynamic characteristics of the robot joint side and the motor side are identified in real time by an optimization algorithm, and the flexible joint model of the robot is updated based on the equivalent rigid inertia correction theory to obtain a joint dynamics equation fused with a closed-loop torque control law, and the joint parameters of the output robot are obtained, which include the angle, acceleration, and torque of the joint side;

[0008] The particle swarm optimization algorithm is used to identify the parameters of the contact model combined with the sensor data of the robot, and a coupled contact model based on the non-elastic impact contact model optimization is constructed to obtain the output normal impact force and tangential friction force;

[0009] Based on the normal impact force, the tangential friction force and the joint parameters, a phase-coupling analysis of continuous contact phase and impact phase is performed using a phase-coupling modeling strategy, and an alternating impact dynamics model is constructed;

[0010] Based on the alternating impact dynamics model, the non-smooth mechanics theory, impact dynamics and task constraints are combined to construct a coupling equation, and the impact dynamics equation is reconstructed by friction correction Jacobian matrix to obtain the post-impact velocity;

[0011] Based on the post-impact velocity, an impact force control strategy is generated to control the work robot to perform alternating impact contact tasks.

[0012] The reconstructed impact dynamics equation is used to describe the velocity variation characteristics of the robot during impact.

[0013] Optionally, the dynamic characteristics of the robot joint side and motor side are identified in real time by an optimization algorithm, and the flexible joint model of the robot is updated based on the equivalent rigid inertia correction theory to obtain a joint dynamics equation fused with closed-loop torque control law, and the output joint parameters of the robot are obtained, including:

[0014] The joint stiffness and motor inertia are identified in real time by frequency domain response data and an optimization algorithm by analyzing the joint side and motor side dynamic characteristics during the impact of the robot.

[0015] Based on the joint stiffness and the motor inertia, a joint dynamics equation fused with closed-loop torque control law is constructed by the equivalent rigid inertia correction theory based on the flexible joint model of the robot, and the output joint parameters of the robot are obtained.

[0016] Optionally, based on the joint stiffness and the motor inertia, a joint dynamics equation fused with closed-loop torque control law is constructed by the equivalent rigid inertia correction theory, and the output joint parameters of the robot are obtained, including:

[0017] The joint stiffness and the motor inertia are input, and based on the equivalent rigid inertia correction theory, the joint dynamics equation is constructed according to to obtain the joint side angle , acceleration and torque in real time.

[0018] Wherein, represents the externally applied joint torque, Indicates the angle on the motor side. Indicates the angular velocity of the joint side. This represents the amount of inertia in the joint space. This indicates the Coriolis, centrifugal, and gravitational effects of the system. Indicates the desired torque on the joint side. This indicates the inertia of the motor.

[0019] Optionally, a particle swarm optimization algorithm combined with the robot's sensor data is used to identify the parameters of the contact model, and a coupled contact model based on the inelastic impact contact model is constructed to obtain the output normal impact force and tangential friction force, including:

[0020] The robot's sensors collect impact contact force data in real time, and the contact force data is identified in the contact model through a particle swarm optimization algorithm to eliminate non-physical viscous forces and optimize tangential velocity errors, thereby obtaining the contact parameters of the contact model, which include contact stiffness, contact damping, and contact model coefficients.

[0021] A coupled contact model based on an inelastic impact contact model optimization is constructed, and the output normal impact force is obtained based on the contact parameters.

[0022] The normal impact force is described by a rate-dependent damping term, and the friction force is described by a modified Coulomb friction term to obtain the tangential friction force.

[0023] Optionally, a coupled contact model optimized based on an inelastic impact contact model is constructed, and the output normal impact force is obtained based on the contact parameters, including:

[0024] Based on contact stiffness Contact damping and contact model coefficients ,according to A coupled contact model based on an inelastic impact contact model was constructed to obtain the output normal impact force. ;

[0025] in, Indicates contact deformation. This indicates the contact deformation rate.

[0026] Optionally, the normal impact force is described based on a rate-dependent damping term, and the frictional force is described by a modified Coulomb friction term to obtain the tangential frictional force, including:

[0027] With normal impact force For input, according to Obtain tangential friction force ;

[0028] in, Indicates contact deformation. Indicates the contact deformation rate. Indicates the coefficient of friction. Indicates impact contact and Resultant velocity in the direction, This represents solving for the magnitude of a vector.

[0029] Optionally, based on the normal impact force, the tangential friction force, and the joint parameters, a phase-separated modeling strategy is used to couple and analyze the continuous contact phase and the impact phase, constructing an alternating impact dynamics model, including:

[0030] Based on normal impact force By employing a phase-separation modeling strategy, the mapping relationship between normal impact force and joint torque is obtained through the dynamic coupling of continuous contact force in flexible joints. ;

[0031] Based on tangential friction By employing a phase-separation modeling strategy and analyzing the velocity mutation through the impulse-momentum equation, the mapping relationship between tangential friction and joint torque is obtained. ;

[0032] Based on mapping relationship and mapping relationship ,according to Construct an alternating impact dynamics model;

[0033] in, , Indicates the joint velocity after impact contact. This indicates the joint velocity before impact contact. Represents the normal Jacobian matrix. This represents the tangential Jacobian matrix.

[0034] Optionally, based on the alternating impact dynamics model, a coupled equation is constructed by combining non-smooth mechanics theory with impact dynamics and task constraints. The impact dynamics equation is then reconstructed using a friction-corrected Jacobian matrix to obtain the post-impact velocity, including:

[0035] Based on the alternating impact dynamics model, a multi-space dynamic coupling framework is constructed by combining non-smooth mechanics theory and impact dynamics.

[0036] Based on the aforementioned multi-space dynamic coupling framework and the Jacobian mapping constrained by task, according to Construct coupling equations that include joints, contacts, and task space;

[0037] Correcting the Jacobian matrix through friction Update the coupling equations according to reconstructing an impact dynamics equation to obtain a post-impact speed ;

[0038] wherein, denotes a unit matrix.

[0039] Optionally, after obtaining the post-impact speed, the method further comprises:

[0040] performing closed-loop checking on the alternating impact dynamics model through multi-source data fusion, and collecting a posture, a speed and a contact force of the robot during the impact process;

[0041] extracting a speed mutation feature after impact release from the posture, the speed and the contact force by polynomial fitting to obtain a post-impact speed, and generating measured data;

[0042] establishing a multi-body dynamics simulation model, calculating a prediction error based on the measured data, and analyzing multi-space stability of the system through joint Lyapunov index to obtain an accuracy evaluation result of the alternating impact dynamics model.

[0043] In a second aspect, the application provides an alternating impact disassembly operation robot dynamic characteristic modeling system, comprising:

[0044] a joint dynamics equation construction module, configured to identify dynamic characteristics of a robot joint side and a motor side in real time through an optimization algorithm, update a flexible joint model of the robot based on an equivalent rigid inertia correction theory, obtain a joint dynamics equation fused with a closed-loop torque control law, and obtain output joint parameters of the robot, the joint parameters including an angle, an acceleration and a torque of the joint side;

[0045] a coupled contact model construction module, configured to identify parameters of a contact model by adopting a particle swarm optimization algorithm in combination with sensor data of the robot, construct a coupled contact model optimized based on a non-elastic impact contact model, and obtain output normal impact force and tangential friction force;

[0046] an alternating impact dynamics model construction module, configured to adopt a phase-by-phase modeling strategy to couple and analyze a continuous contact phase and an impact phase based on the normal impact force, the tangential friction force and the joint parameters, and construct an alternating impact dynamics model;

[0047] an impact dynamics equation reconstruction module, configured to, on the basis of the alternating impact dynamics model, construct a coupled equation in combination with non-smooth mechanics theory, impact dynamics and task constraints, and reconstruct an impact dynamics equation by friction correction Jacobian matrix to obtain a post-impact speed;

[0048] an impact force control module, configured to generate an impact force control strategy based on the post-impact speed, and control the operation robot to perform an alternating impact contact task.

[0049] wherein the reconstructed impact dynamics equation is used to describe the velocity change characteristics in the robot impact process.

[0050] To sum up, in order to realize the construction of an accurate impact dynamics model, first, the dynamic characteristics of the robot joint side and the motor side are comprehensively considered, a torque-driven flexible joint dynamics model is established, and the influence of joint flexibility on impact response is determined. Then, the exponential inelastic continuous contact model is used to estimate the radial impact force and the tangential friction force, and an alternating impact dynamics model is constructed to quantify the momentum change and nonlinear dynamics effects in the impact process. Finally, a multi-space dynamic characteristic coupling model is established to systematically analyze the interaction mechanism of joint space, impact contact space and task space, output accurate post-impact velocity, generate impact force control strategy, improve the dynamic response accuracy of the robot under alternating impact operation, realize the active use of the robot to impact, and ensure the stability, accuracy and efficiency of the disassembly operation. As can be seen, the present application not only enriches and perfects the theoretical system of robotized manufacturing, but also expands it from “non-contact” and “soft contact” to “impact contact” category, and provides a solid theoretical support and technical guarantee for the application of robots in efficient disassembly, complex assembly and high dynamic operation environment, which has important scientific research value and application prospect, and effectively solves the problem of insufficient impact force control in disassembly operation due to the lack of accurate impact dynamics model in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0051] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0053] Figure 1 A flowchart of a dynamic characteristic modeling method of an alternating impact disassembly operation robot provided by an embodiment of the present application;

[0054] Figure 2 A step flowchart of a dynamic characteristic modeling method of an alternating impact disassembly operation robot provided by an optional embodiment of the present application;

[0055] Figure 3 A dynamic characteristic modeling diagram of an impact disassembly operation robot provided by an embodiment of the present application;

[0056] Figure 4 is a block diagram of a dynamic characteristic modeling system of an alternating impact disassembly operation robot provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in a clear and complete manner with reference to the drawings accompanying the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in a clear and complete manner with reference to the drawings accompanying the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0059] Figure 1 A flowchart of a method provided by an embodiment of the present application is shown. As shown in the figure, the alternating impact disassembly operation robot dynamic characteristic modeling method provided by the embodiment of the present application can specifically include the following steps: Figure 1

[0060] In step 110, the dynamic characteristics of the robot joint side and the motor side are identified in real time through an optimization algorithm, and the flexible joint model of the robot is updated based on the equivalent rigid inertia correction theory to obtain a joint dynamics equation fused with a closed-loop torque control law, and the joint parameters of the output robot are acquired.

[0061] The joint parameters include the angle, acceleration, and torque of the joint side.

[0062] In the embodiment, the optimization algorithm for identifying the dynamic characteristics in real time can include but is not limited to a particle swarm optimization algorithm (PSO), an adaptive optimization algorithm, and the like.

[0063] In the related art, the existing alternating impact disassembly operation robot faces multiple challenges in actual application. For example, the impact robot flexible joint model lacks consideration of the influence of the low-order torque control loop. Specifically, since the robot joint usually has a flexible characteristic, its dynamic behavior is affected by the spring-damper unit, and under the action of high-frequency impact, elastic oscillation may be generated, affecting the transmission efficiency and accuracy of the force.

[0064] ​To solve the problem, the embodiment improves the existing flexible joint model of the impact robot, updates the existing flexible joint model by using the equivalent rigid inertia correction theory, constructs an updated joint dynamics model, and obtains a joint dynamics equation fused with a closed-loop torque control law. The joint dynamics equation comprehensively considers the dynamic characteristics of the joint side and the motor side, takes the quantified dynamic characteristics such as joint stiffness and motor inertia as inputs, obtains the joint torque, joint side angular acceleration, and joint side angle of the robot during the alternating impact disassembly operation, and takes the joint parameters as inputs.

[0065] In actual implementation, to introduce the dynamic characteristics of the joint side and the motor side, the embodiment mainly identifies the dynamic characteristics of the joint side and the motor side of the robot in real time through frequency domain response data and an optimization algorithm, and obtains the joint stiffness and motor inertia (or motor inertia). Specifically, force sensors and the like are preset in the robot, the response data (such as a response curve) of the motor side angle and the joint side angle are collected, and the response data of the joint torque and the externally applied torque are synchronously collected. The collected data are processed by Fourier transform to obtain frequency domain response data, which reflects the dynamic characteristics (such as the resonance frequency and the damping ratio) of the joint at different frequencies. Then, the frequency domain response data are iteratively calculated by using the optimization algorithm, so that the error between the joint parameters output by the model and the measured values is minimized, the model is updated, the speed jump prediction error of the traditional model is significantly reduced, and the optimal parameters output by the model are obtained.

[0066] It should be noted that the robot in the embodiment is used to perform impact disassembly operation, and therefore, the robot can refer to an impact disassembly operation robot, which is referred to as a robot or an operation robot, and the like. The embodiment does not limit this.

[0067] In step 120, the particle swarm optimization algorithm is used to identify the parameters of the contact model in combination with the sensor data of the robot, a coupled contact model optimized based on the non-elastic impact contact model is constructed, and the output normal impact force and tangential friction force are obtained.

[0068] In the embodiment, the sensor data includes force sensor data and torque sensor data; and the coupled contact model refers to an exponential normal-tangential coupled contact model.

[0069] In specific implementation, considering that the traditional non-elastic impact contact model only focuses on the normal impact force and ignores the dynamic coupling effect of the tangential friction force, resulting in non-physical phenomena such as impact force mutation and contact force stickiness, it is difficult to represent the energy dissipation mechanism in the actual impact process. The embodiment improves the traditional non-elastic impact contact model, constructs an exponential normal-tangential coupled contact model, identifies the contact model parameters in combination with the force / torque sensor data by using the particle swarm optimization algorithm, eliminates the non-physical stickiness force and reduces the tangential velocity error, and thereby obtains the output normal impact force and tangential friction force.

[0070] In actual implementation, the joint parameters output by the joint dynamics equation and the input parameters of the coupled contact model are closely related, and the two are influenced by each other through the physical coupling relationship of the robot joint movement and the impact contact process.

[0071] In step 130, based on the normal impact force, the tangential friction force and the joint parameters, a phase-coupling modeling strategy is used to analyze the continuous contact phase and the impact phase, and an alternating impact dynamics model is constructed.

[0072] In specific implementation, the impact contact process involves dynamic coupling of the normal impact force and the tangential friction force, but under high-frequency alternating impact, the coupling relationship has strong nonlinear characteristics, and the rigid contact model in the existing disassembly robot cannot accurately describe it. Therefore, the embodiment adopts a phase-coupling modeling strategy to couple the contact phase and the impact phase, wherein the continuous contact phase is coupled with the contact force through the flexible joint dynamics, and the impact phase is based on the impulse-momentum equation to analyze the velocity jump. Thus, the embodiment couples the normal impact force and the tangential friction force estimated by the exponential inelastic continuous contact model with the joint parameters output by the flexible joint model, to construct an alternating impact dynamics model, and to realize the quantification of the momentum change and the nonlinear dynamics effect in the impact process.

[0073] In step 140, based on the alternating impact dynamics model, the non-smooth mechanics theory, impact dynamics and task constraints are combined to construct a coupling equation, and the impact dynamics equation is reconstructed by modifying the Jacobian matrix, to obtain the velocity after impact.

[0074] The reconstructed impact dynamics equation is used to describe the velocity change characteristics of the robot during impact.

[0075] In specific implementation, considering that during the robot task execution process, complex mapping needs to be performed between the joint space, the impact contact space and the task space, this multi-space dynamics coupling effect further increases the difficulty of modeling and control. Therefore, the embodiment analyzes the multi-space dynamic characteristic coupling mechanism in the impact contact process, including systematically analyzing the interaction mechanism of the joint space, the impact contact space and the task space, to construct an accurate dynamics model.

[0076] Specifically, based on the alternating impact dynamics model, first, the multi-space dynamic coupling framework is established by combining non-smooth mechanics theory and impact dynamics, so as to realize the coupling of joint space, impact contact space and task space. Then, combined with task constraints (such as Jacobian mapping, etc.), the coupling equations of multi-space are constructed. Further, in the coupling equations, considering the characteristics of different spaces, the Jacobian matrix is modified by friction to build a bridge between multi-spaces, realizing the coupling of multi-spaces, such as the bridge between robot joint space and task space. Finally, based on the impulse-momentum theorem and the multi-space dynamics coupling theory, the post-impact velocity is derived, and the impact dynamics equation can accurately describe the sudden change relationship of joint velocity in the impact contact process.

[0077] At step 150, based on the post-impact velocity, an impact force control strategy is generated, and the work robot is controlled to perform the alternating impact contact task.

[0078] In a specific implementation, when the robot performs the alternating impact contact task (i.e., impact disassembly work), an impact force control strategy is generated according to the post-impact velocity to achieve precise dynamic control of the robot in the alternating impact environment.

[0079] It can be seen that the embodiments of the present application aim to break through the theoretical bottleneck of existing robot control technology in alternating impact contact work, realize active regulation of impact force, and improve the adaptability and work precision of robots in complex disassembly tasks. Specifically, for the existing robots in handling alternating impact tasks, they often rely on passive buffering or rigid control strategies, which are difficult to effectively cope with the coupling effects of non-elastic impact, joint flexible deformation and task constraints, etc., resulting in problems such as impact force overrun, uncertainty sliding and trajectory deviation, which further limits their application in the field of fine disassembly. The embodiments of the present application propose a dynamic characteristic modeling method for alternating impact disassembly work robots, which systematically analyzes the multi-space dynamic characteristic coupling mechanism in the impact contact process, constructs an accurate dynamics model, and introduces an impact force control strategy to realize active utilization of impact by the robot, ensuring the stability, accuracy and efficiency of the disassembly work. Not only does it enrich and perfect the theoretical system of robotized manufacturing, but also expands it from "non-contact" and "soft contact" to "impact contact" category, and provides a solid theoretical support and technical guarantee for the application of robots in efficient disassembly, complex assembly and high dynamic work environment, which has important scientific research value and application prospect.

[0080] Referring to Figure 2 FIG. 1 shows a step flow diagram of an alternating impact disassembly work robot dynamic characteristic modeling method according to an optional embodiment of the present application. The method can specifically include the following steps:

[0081] At step 210, the dynamic characteristics of the robot joint side and the motor side are identified in real time through an optimization algorithm, and the flexible joint model of the robot is updated based on the equivalent rigid inertia correction theory to obtain a joint dynamics equation of the fusion closed-loop torque control law, and the joint parameters of the output robot are obtained.

[0082] The joint parameters include the angle, acceleration and torque of the joint side.

[0083] The description of step 210 can refer to step 110 described above, and this embodiment will not be described in detail.

[0084] Optionally, the above-mentioned identification of the dynamic characteristics of the robot joint side and the motor side in real time through the optimization algorithm, and the updating of the flexible joint model of the robot based on the equivalent rigid inertia correction theory to obtain the joint dynamics equation of the fusion closed-loop torque control law, and the output of the joint parameters of the robot can include the following sub-steps:

[0085] At sub-step 2101, the dynamic characteristics of the joint side and the motor side during the impact of the robot are analyzed, and the joint stiffness and the motor inertia are identified in real time through the frequency domain response data and the optimization algorithm.

[0086] At sub-step 2102, based on the joint stiffness and the motor inertia, the joint dynamics equation of the fusion closed-loop torque control law is constructed through the equivalent rigid inertia correction theory based on the flexible joint model of the robot, and the joint parameters of the output robot are output.

[0087] The sub-steps 2101-2102 are described uniformly as follows:

[0088] In specific implementation, considering that the existing flexible joint model of the impact robot only considers the dynamic characteristics of the joint side (such as a linear spring-damper system), and ignores the influence of the low-order torque control loop of the motor side (such as motor inertia, transmission ratio, closed-loop control strategy, etc.), it is easy to cause various problems, including: large model error: the nonlinear correction of the motor control to the joint flexibility is not reflected, and the speed jump under high-frequency impact cannot be accurately described; insufficient prediction accuracy: under high-speed impact conditions, the actual impact of the motor dynamic characteristics on the force transmission cannot be reflected, resulting in a significant deviation between the dynamic model and the actual working condition.

[0089] Therefore, the flexible joint model in this embodiment is an improvement on existing models, comprehensively considering the bidirectional coupling effect of the joint side (mechanical structural flexibility) and the motor side (dynamic characteristics of the control loop). First, the dynamic characteristics of the joint side and motor side during the alternating impact disassembly task are analyzed. This is achieved through a frequency domain response data method: using frequency domain response data (such as the vibration response of the joint at different frequencies) and an optimization algorithm to dynamically identify joint stiffness and motor inertia. The goal of this optimization algorithm is to correct the model parameters to more closely approximate the actual robot dynamics, reducing the speed jump prediction error caused by parameter mismatch in traditional models (such as the calculation deviation of sudden speed changes before and after impact). Then, based on the equivalent rigid inertia correction theory, by introducing parameters such as motor inertia, the motor control characteristics are transformed into correction terms for joint dynamics, constructing a more realistic flexible joint model. Furthermore, a joint dynamics equation integrating a closed-loop torque control law is constructed, outputting the robot's joint parameters.

[0090] In one optional embodiment, this embodiment, based on the joint stiffness and the motor inertia, constructs a joint dynamics equation that integrates a closed-loop torque control law through the equivalent rigid inertia correction theory. The output joint parameters of the robot can specifically include: joint stiffness... and motor inertia As input, based on the equivalent rigid inertia correction theory, according to Construct joint dynamic equations to obtain joint angles in real time. acceleration and torque ;in, This indicates the joint torque applied externally. Indicates the angle on the motor side. Indicates the angular velocity of the joint side. This represents the amount of inertia in the joint space. This indicates the Coriolis, centrifugal, and gravitational effects of the system. Indicates the desired torque on the joint side. This indicates the inertia of the motor.

[0091] Reference Figure 3 As shown, the dynamic equations of the fusion closed-loop torque control law established in this embodiment are mainly divided into two parts, namely the torque balance equations. and equations of motion Among them, the torque balance equation, driven by the joint torque and the difference between the motor side angle and the joint side angle, reflects the resistance of joint stiffness to flexible deformation, describes the elastic torque generated by joint flexibility, and reflects the deformation characteristics of the mechanical structure; the motion equation is mainly used to reflect inertial forces. (Joint inertia) With motor inertia sum of equivalent inertias), coriolis force , centrifugal force and gravity, by integrating motor-side inertia and joint-side inertia, the direct influence of motor dynamic characteristics on joint motion is embodied, and the defects of traditional models ignoring motor control loop are solved.

[0092] Therefore, by introducing motor-side dynamic characteristics, the model can more accurately describe the coupling effect of joint flexibility and motor control, and reduce the dynamic error caused by ignoring the control loop. The model breaks through the limitations of traditional linear models by integrating the dynamic characteristics of "joint-motor" on both sides, realizes accurate modeling of the nonlinear behavior of the robot flexible joint during impact, and effectively improves the model accuracy. It is the basis for subsequent alternating impact dynamics and multi-space coupling models.

[0093] In addition, the embodiment provides a more reliable theoretical basis for the impact force control of the alternating impact robot disassembly machine, improves the stability and trajectory accuracy of the robot in high-frequency impact operation, and is especially suitable for scenes such as disassembly of retired power batteries that require precise force control.

[0094] Step 220, real-time acquisition of impact contact force data based on the sensor of the robot, and identification of the contact force data in the contact model through a particle swarm optimization algorithm to eliminate non-physical viscous force and optimize tangential velocity error, and obtain contact parameters of the contact model.

[0095] The contact parameters include contact stiffness, contact damping, and contact model coefficients.

[0096] Step 230, constructing a coupled contact model based on the optimization of the inelastic impact contact model, and obtaining the output normal impact force based on the contact parameters.

[0097] The steps 220-230 are uniformly described as follows:

[0098] In related technologies, existing robot impact contact models are typically inelastic impact contact models, which only consider the normal impact force (the force perpendicular to the contact surface) and ignore the dynamic coupling effect of the tangential friction force (the force parallel to the contact surface). This easily leads to: non-physical phenomena: "abrupt changes in impact force" (such as a jump in force value at the moment of contact) and "contact force stickiness" (such as the force not returning to zero in time after separation) in the model predictions, which do not match the actual impact process; and failure to describe energy dissipation: the model cannot characterize the energy loss caused by friction during the impact process (such as heat and sliding loss), resulting in the model's inability to accurately reflect the real dynamic behavior. In addition, since the impact contact process involves the dynamic coupling of the normal impact force and the tangential friction force, especially under high-frequency alternating impact, this coupling relationship has strong nonlinear characteristics and is difficult to accurately describe using existing rigid contact models. In practical scenarios such as the dismantling of retired power batteries, impact contact is often accompanied by sliding (such as the relative motion between tools and rusted components), and the tangential friction force has a significant impact on the impact force transmission, dismantling efficiency, and tool wear. Traditional models cannot meet engineering requirements.

[0099] This embodiment improves upon the shortcomings of traditional contact models by proposing an exponential normal-tangential coupled contact model. This coupled contact model introduces a dynamic coupling mechanism between normal and tangential forces, describes the nonlinear characteristics of the impact process through an exponential function and a modified friction term, and optimizes model parameters using sensor data to improve the accuracy of contact force prediction.

[0100] Specifically, this embodiment first uses a particle swarm optimization algorithm to identify the impact contact force data collected in real time by the robot's sensors, in order to eliminate non-physical viscous forces and optimize tangential velocity errors, thereby obtaining the contact parameters of the contact model. Then, in the constructed coupled contact model, the normal impact force is identified in real time using the contact parameters as input.

[0101] In an optional embodiment, the above-described construction of a coupled contact model optimized based on an inelastic impact contact model, and the acquisition of the output normal impact force based on the contact parameters, may include: based on contact stiffness. Contact damping and contact model coefficients ,according to A coupled contact model based on an inelastic impact contact model was constructed to obtain the output normal impact force. ;in, Indicates contact deformation. This indicates the contact deformation rate.

[0102] Reference Figure 3 As shown, in this embodiment, the impact normal force is composed of the elastic term. and damping term The two together reflect the elastic deformation and the rate-dependent energy dissipation existing simultaneously in the impact process.

[0103] To be closer to the actual contact and separation process, a segmented function is used to describe the normal impact force under different contact states . Specifically, a segmented function is used to describe the attenuation of force with deformation rate to avoid the sudden change of force in traditional models. or , it indicates that it is in the contact state and in the impact contact phase (at this time, the deformation increases), at this time, the normal force is obtained according to ; when and , it indicates that it is in the contact state and in the contact separation phase (at this time, the deformation decreases), at this time, the normal force is obtained according to ; when , it indicates that it is in the non-contact state, i.e. in the non-contact phase, at this time, the normal force is obtained according to .

[0104] In this embodiment, the deformation rate term is introduced to reflect the influence of impact speed on force (such as the increase of damping effect at high speed impact); and the contact model coefficient is introduced to adjust the nonlinearity degree of force-deformation relationship.

[0105] Therefore, this embodiment eliminates the non-physical viscous force in the existing impact contact model, improves the coupling accuracy of normal impact force and tangential friction force. Compared with the existing model, the present application is closer to the energy dissipation mechanism in the actual impact process, and effectively reduces the impact speed prediction error.

[0106] In step 240, the normal impact force is described based on the rate-dependent damping term, and the friction force is described by combining the modified Coulomb friction term to obtain the tangential friction force.

[0107] This embodiment breaks through the limitation of the traditional rigid contact model, and incorporates the tangential friction into the dynamic analysis framework. Specifically, based on the coupled contact model, the friction force is described by combining the modified Coulomb friction term, the tangential friction force model is constructed, and it is coupled into the contact model to form an exponential normal-tangential coupling contact model, which provides more perfect theoretical support for the application of robots in high-frequency alternating impact scenarios.

[0108] Optionally, the normal impact force is described based on the rate-dependent damping term, and the friction force is described by combining the modified Coulomb friction term to obtain the tangential friction force, which can specifically include: taking the normal impact force as input, and obtaining the tangential friction force according to ; wherein represents the contact deformation, represents the contact deformation rate, denotes the friction coefficient, denotes the impact contact and the resultant velocity of the direction, denotes the vector module solution.

[0109] In the embodiment, the resultant velocity reflects the relative sliding direction and speed size of the contact point, and is a unit speed vector, for ensuring that the friction force direction is always opposite to the sliding direction. Considering that the normal force is the premise of generating tangential friction, the two have a coupling relationship. In the impact scenario, the normal impact force directly determines the degree of extrusion between the contact surfaces, and then affects the size of the friction force. For example, the greater the normal impact force, the greater the friction resistance between the contact surfaces, that is, the tangential friction force is proportional to the normal impact force, and its direction is opposite to the direction of the relative sliding speed. Therefore, the embodiment incorporates the tangential friction effect into the impact contact model to calculate the tangential friction force and solve the defects of the traditional model that ignores the friction coupling.

[0110] In actual implementation, the embodiment adopts a particle swarm optimization algorithm to combine the measured data of the force / torque sensor to jointly identify model parameters (including: , , and ). The goal is to eliminate the "non-physical viscous force" (such as unreasonable friction force remaining after separation) caused by parameter mismatch in the traditional model and reduce the tangential velocity prediction error (such as sliding speed calculation deviation).

[0111] In specific implementation, the joint dynamics equation provides a motion basis for the coupled contact model, and the joint torque and joint side parameters solved by the joint dynamics equation have relevance with the input parameters of the coupled contact model, such as the joint side angular velocity is mapped to the contact space through the Jacobian matrix to form the relative velocity of the impact contact. If or is inaccurate, it may cause joint motion prediction error (such as velocity jump), and then make the contact deformation variables and deformation rates calculated in the contact model deviate from the actual values, and then affect the accuracy of the normal force and the friction force. Therefore, the embodiment dynamically identifies and through the frequency domain response data and the optimization algorithm to reflect the influence of joint flexibility on the impact response and obtain accurate parameters.

[0112] It can be seen that in the embodiment, the joint side parameters provide motion boundary conditions for the coupled contact model; the force output of the coupled contact model is fed back as external excitation to the joint dynamics equation, affecting the subsequent joint torque and motion state. This coupling relationship is the core of realizing precise modeling and control of the impact disassembly robot, and through collaborative analysis of multi-space dynamic characteristics, the impact force prediction and control accuracy is significantly improved.

[0113] Therefore, the embodiment adopts a more realistic physical characterization, accurately describes the energy dissipation mechanism in the impact process through normal-tangential coupling modeling, avoids non-physical phenomena such as “force jump” and “viscous force”, and improves the consistency of the model with the actual working condition.

[0114] In step 250, based on the normal impact force, the tangential friction force and the joint parameters, a phase separation modeling strategy is adopted to analyze the continuous contact phase and the impact phase, and an alternating impact dynamics model is constructed.

[0115] In the embodiment, based on the coupled contact model, the invention adopts a phase separation modeling strategy to analyze the continuous contact phase and the impact phase, the continuous contact phase couples the contact force through the flexible joint dynamics, and the impact phase analyzes the velocity jump based on the impulse-momentum equation. In this way, an alternating impact dynamics model is constructed.

[0116] In an optional embodiment, based on the normal impact force, the tangential friction force and the joint parameters, the invention adopts a phase separation modeling strategy to analyze the continuous contact phase and the impact phase, and an alternating impact dynamics model is constructed, which can include: based on the normal impact force , a phase separation modeling strategy is adopted to couple the continuous contact force through the flexible joint dynamics, and the mapping relationship between the normal impact force and the joint torque is obtained ; based on the tangential friction force , a phase separation modeling strategy is adopted to analyze the velocity jump through the impulse-momentum equation, and the mapping relationship between the tangential friction force and the joint torque is obtained ; based on the mapping relationship and the mapping relationship , according to , an alternating impact dynamics model is constructed; wherein, , represents the joint velocity after impact contact, represents the joint velocity before impact contact, represents the normal Jacobian matrix, represents the tangential Jacobian matrix.

[0117] Referring to Figure 3 , in the embodiment, the normal impact force and the tangential friction force calculated by the coupled contact model are analyzed through the impulse-momentum equation , an alternating impact dynamics model is constructed to describe the dynamics coupling relationship among joint space, contact space and task space, so as to reflect the dynamic influence of impact contact force on robot joint motion. In the impulse-momentum equation, the input force is converted into external torque in the joint space , which is one of the input parameters of the joint dynamics equation and affects the subsequent acceleration and torque output of joint motion.

[0118] In the impulse-momentum equation, the initial velocity after impact is obtained by solving the velocity jump: , and the continuous contact phase takes as the initial condition to calculate the continuous influence (such as acceleration, torque change, etc.) of contact force on joint motion through flexible joint dynamics equation. Through loop iteration, when the contact is separated, i.e. , the model is reset to the non-contact state; if it is contacted again, the sub-phase calculation of “impact phase-continuous contact phase” is repeated.

[0119] In the specific implementation, the and generated in the impact contact process are mapped to the joint space through Jacobian matrices and , change the dynamic response of the inertia and motor inertia on the joint side, and thus correct the joint torque and angle , and construct the corresponding mapping relationship (i.e. impulse of normal force and tangential friction), i.e. and .

[0120] Therefore, the embodiment jointly acts on the multi-space coupling equation and the joint side parameters to correct the acceleration and torque of joint motion, forming a closed loop coupling of “joint motion-contact force calculation-torque feedback”. It mainly realizes: ① separate transient and steady state characteristics: the impact phase focuses on the velocity jump, and the continuous phase focuses on the force-motion coupling, which improves the accuracy of the model in a wide time scale; ② multi-physical field mapping: dynamic mapping between contact space and joint space is realized through Jacobian matrix, solving the multi-space coupling problem; ③ parameter reusability: joint side parameters and contact model parameters can be optimized through the same set of identification process (such as frequency domain response, force sensor data), ensuring the consistency of the model.

[0121] Therefore, for the dynamic coupling of the normal impact force and the tangential friction force involved in the impact contact process, especially under high-frequency alternating impact, the coupling relationship has strong nonlinear characteristics, and it is difficult to accurately describe it by using the traditional rigid contact model. In this embodiment, a phase separation modeling strategy is adopted to solve the modeling contradiction between "transient speed mutation" and "steady force coupling" in the alternating impact process through physical process decomposition and mathematical tool adaptation. The impulse-momentum equation of the impact phase and the flexible joint dynamics equation of the continuous contact phase form a complement, which together builds a dynamic model that accurately describes the impact response of the robot, providing a theoretical basis for impact force control and operation stability improvement.

[0122] In step 260, on the basis of the alternating impact dynamics model, a coupling equation is constructed by combining the non-smooth mechanics theory and impact dynamics, task constraints, and the impact dynamics equation is reconstructed by modifying the Jacobian matrix with friction to obtain the post-impact speed.

[0123] The reconstructed impact dynamics equation is used to describe the speed change characteristics of the robot during the impact process.

[0124] In a specific implementation, on the basis of the alternating impact dynamics model, a multi-space dynamic coupling model is established to systematically analyze the interaction mechanism of the joint space, the impact contact space, and the task space. This multi-space dynamic coupling analysis framework mainly analyzes the speed mutation after impact based on the non-smooth mechanics theory and the impulse-momentum equation, and accurately predicts the post-impact speed by combining the friction-modified Jacobian matrix. Compared with traditional dynamic modeling methods, this invention can systematically reveal the synergistic effect between the robot's flexible joint, contact nonlinearity, and task constraints, laying a solid theoretical and technical foundation for high-precision control and stable operation of alternating impact operating robots. It not only helps to improve the automation level of retired power battery disassembly, but also provides a new idea for impact dynamics modeling and control.

[0125] Optionally, on the basis of the alternating impact dynamics model, a coupling equation is constructed by combining the non-smooth mechanics theory and impact dynamics, task constraints, and the impact dynamics equation is reconstructed by modifying the Jacobian matrix with friction to obtain the post-impact speed, including: on the basis of the alternating impact dynamics model, a multi-space dynamic coupling framework is constructed by combining the non-smooth mechanics theory and impact dynamics; based on the Jacobian mapping of the multi-space dynamic coupling framework and the task constraints, the coupling equation containing the joint, contact, and task spaces is constructed according to , the coupling equation is updated by modifying the Jacobian matrix with friction , the impact dynamics equation is reconstructed according to , and the post-impact speed is obtained; wherein I represents the identity matrix.

[0126] In the multi-space dynamic coupling implemented in the embodiment, the multi-space mainly includes joint space, contact space and task space (guided by disassembly trajectory and work target, joint motion is mapped to the pose of tool end through Jacobian matrix).

[0127] Referring to Figure 3 , firstly, the embodiment analyzes the joint momentum after impact, the joint momentum before impact and the impulse of contact force before impact, and based on the impulse conservation law of non-smooth mechanics, describes that the change of joint momentum at the moment of impact is directly determined by the impulse of contact force, skips the intermediate state of impact process, and simplifies the calculation. By combining the Jacobian mapping of task constraints, a multi-space dynamic coupling equation including joint, contact and task spaces is constructed: . In the coupling equation, Jacobian matrix is used to associate and map different spaces. Specifically, through the Jacobian matrix, the trajectory requirements of disassembly work (such as tool impact in a specific direction) are converted into motion constraints in joint space, ensuring that the robot action meets the task target.

[0128] Then, considering that the traditional Jacobian matrix only describes the mapping of normal motion and joint, but the tangential friction force will change the transmission path of force in actual impact, therefore, the embodiment introduces a friction modified Jacobian matrix , which introduces term in the matrix to correct the coupling effect of friction on normal force transmission, so that the Jacobian matrix contains the joint action of normal-tangential force, and more truly maps the relationship between contact force and joint motion.

[0129] Finally, the influence of friction on the velocity after impact is quantified, the velocity prediction deviation caused by ignoring tangential coupling in the traditional model is avoided (such as the reduction of tangential velocity error), and the integrated solution of multi-space parameters is realized through matrix operation, which improves the calculation efficiency and stability of the model. Specifically, based on the impulse-momentum equation and the friction modified Jacobian matrix, the velocity increment after impact is separated out through matrix inverse operation, and the velocity after impact is obtained. The velocity after impact formula is obtained through term to represent the correction effect of contact force impulse on the original velocity; the unit matrix is used to ensure that the velocity after impact contains the original velocity component and the correction component.

[0130] In actual implementation, the coupling equation and the separate modeling strategy have a synergistic effect, mainly reflected in: ①, the multi-space coupling process of impact phase, the Jacobian mapping converts the normal force and friction force into the impulse of joint space, and the velocity after impact is calculated as the initial condition of continuous contact phase; ②, the velocity The initial velocity is obtained, and subsequent motion is calculated through the flexible joint dynamics equation to realize smooth transition from the impact transient state to the continuous dynamic state.

[0131] Thus, the embodiment breaks through the limitation of the traditional model and realizes accurate description of non-smooth events. First, impact dynamics is used to avoid the problem of invalid differential equation at the impact moment, improving the modeling accuracy under high-frequency impact. Second, the systematicity of multi-physical field coupling is realized, integrating joint flexibility, contact nonlinearity and task constraints (Jacobian mapping), solving the error accumulation caused by traditional separate modeling. Finally, the explicit modeling of friction effect, the friction correction Jacobian matrix first introduces the tangential friction force into the dynamics mapping, which is closer to the actual disassembly scene.

[0132] The multi-space dynamic coupling framework of the embodiment processes the impact transient state characteristics through non-smooth mechanics theory, realizes the linkage of multi-space parameters through Jacobian mapping, and enhances the adaptability of the model to the real contact environment through the friction correction matrix. This technology integrates the traditional separate joint dynamics, contact mechanics and task planning into a unified mathematical system, providing a complete solution from theoretical modeling to engineering implementation for high-precision control of alternating impact robots.

[0133] Further, the post-impact velocity obtained by the embodiment provides a theoretical model for impact force control strategies (such as actively adjusting impact frequency and direction), helping to realize the automatic process of flexible impact, efficient demolition and precise separation.

[0134] Step 270, based on the post-impact velocity, an impact force control strategy is generated, and the robot is controlled to perform an alternating impact contact task.

[0135] Optionally, after obtaining the post-impact velocity, the embodiment can further include: performing closed-loop checking on the alternating impact dynamics model through multi-source data fusion, and collecting the posture, velocity and contact force of the robot during the impact process; using polynomial fitting to extract the velocity mutation characteristics after impact removal from the posture, velocity and contact force to obtain the post-impact velocity, and generating measured data; establishing a multi-body dynamics simulation model, calculating the prediction error based on the measured data, and analyzing the multi-space stability of the system through joint Lyapunov index to obtain the accuracy evaluation result of the alternating impact dynamics model.

[0136] In a specific implementation, in the experimental test phase, the alternating impact dynamics model is closed-loop checked through multi-source data fusion, and high-precision IMU / force sensors are used to collect the pose, speed and contact force during the impact process, and the post-impact speed is extracted by polynomial fitting; by establishing a multi-body dynamics simulation model, the measured data is used as the benchmark to calculate the prediction error (such as normal force, speed jump, etc.), and the multi-space stability of the system is analyzed by joint Lyapunov index to ensure the dynamic consistency of the model. The proposed technical route breaks through the limitations of traditional separate modeling by dynamic parameter correction, impact contact phase modeling, multi-space coupling analysis and experimental-simulation closed-loop verification, and lays a foundation for the dynamic characteristics of the alternating impact disassembly robot.

[0137] In summary, the traditional method does not fully consider the nonlinear influence of motor inertia, transmission ratio and other motor side characteristics on joint flexibility, and ignores the tangential friction effect in the contact space, resulting in inaccurate description of impact speed jump and energy dissipation. In order to realize the construction of an accurate impact dynamics model, a dynamic characteristic speed layer representation and phase-coupled modeling method is proposed as a dynamic characteristic modeling method for alternating impact disassembly robots. This method breaks through the limitations of traditional rigid assumptions and linear flexible models, and constructs a multi-space nonlinear coupled alternating impact dynamics model. First, the dynamic characteristics of the robot joint side and the motor side are considered comprehensively, and based on the equivalent rigid inertia correction theory, a torque-driven flexible joint dynamics model is established to determine the influence of joint flexibility on impact response. Then, the exponential inelastic continuous contact model is used to estimate the normal impact force and tangential friction force, and the normal-tangential coupled modeling is combined to construct the alternating impact dynamics model, which quantifies the momentum change and nonlinear dynamics effect during the impact process. Finally, a multi-space dynamic characteristic coupling model is established, and the task constraint Jacobian matrix is combined to analyze the interaction mechanism of joint space, impact contact space and task space, reconstruct the impact dynamics equation, and output the accurate post-impact speed to accurately describe the speed change characteristics during the impact process. Based on the post-impact speed, an impact force control strategy is generated to improve the dynamic response accuracy of the robot under alternating impact operation, realize the active use of the robot to impact, and ensure the stability, accuracy and efficiency of the disassembly operation.

[0138] Therefore, compared with existing rigid-flexible coupled dynamics methods, the present embodiment reduces the computational complexity to linear logarithmic order while maintaining high accuracy, effectively improves the computational efficiency, and makes the modeling and prediction of impact speed more accurate, providing a theoretical support for the dynamic control of robots in alternating impact environments.

[0139] The embodiment not only enriches and perfects the theoretical system of robotized manufacturing, and expands from "non-contact" and "soft contact" to "impact contact", but also provides solid theoretical support and technical guarantee for the application of robots in efficient disassembly, complex assembly and high dynamic operation environment, has important scientific research value and application prospect, and effectively solves the problem of insufficient impact force control in disassembly operation due to lack of accurate impact dynamics model in the prior art.

[0140] It should be noted that, for the method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the embodiments of the present application are not limited by the action sequence described, because according to the embodiments of the present application, certain steps can be performed in other sequences or simultaneously.

[0141] As shown in Figure 4 The application embodiment also provides an alternating impact disassembly operation robot dynamic characteristic modeling system 400, which comprises:

[0142] The joint dynamics equation construction module 410 is used for identifying the dynamic characteristics of the robot joint side and the motor side in real time through an optimization algorithm, updating the flexible joint model of the robot based on the equivalent rigid inertia correction theory, obtaining the joint dynamics equation fused with the closed-loop torque control law, and acquiring the joint parameters of the output robot, wherein the joint parameters include the angle, acceleration and torque of the joint side;

[0143] The coupled contact model construction module 420 is used for identifying the parameters of the contact model by using a particle swarm optimization algorithm in combination with the sensor data of the robot, constructing a coupled contact model optimized based on a non-elastic impact contact model, and acquiring the normal impact force and tangential friction force as output;

[0144] The alternating impact dynamics model construction module 430 is used for coupling and analyzing the continuous contact phase and the impact phase by using a phase-by-phase modeling strategy based on the normal impact force, the tangential friction force and the joint parameters, and constructing an alternating impact dynamics model;

[0145] The impact dynamics equation reconstruction module 440 is used for constructing a coupled equation by combining the non-smooth mechanics theory with impact dynamics and task constraints on the basis of the alternating impact dynamics model, and reconstructing the impact dynamics equation by friction correction Jacobian matrix, and acquiring the post-impact velocity;

[0146] The impact force control module 450 is used for generating an impact force control strategy based on the post-impact velocity, and controlling the operation robot to perform an alternating impact contact task;

[0147] The reconstructed impact dynamics equation is used for describing the velocity change characteristics of the robot in the impact process.

[0148] Optionally, the joint dynamics equation construction module comprises:

[0149] The dynamic characteristic identification submodule is configured to analyze the joint side and motor side dynamic characteristics in the robot impact process, and identify the joint stiffness and motor inertia in real time through frequency domain response data and an optimization algorithm;

[0150] The joint dynamics equation construction submodule is configured to construct a joint dynamics equation fused with a closed-loop torque control law based on the joint stiffness and the motor inertia through an equivalent rigid inertia correction theory on the basis of the robot flexible joint model, and output joint parameters of the robot.

[0151] Optionally, the coupling contact model construction module comprises:

[0152] The contact parameter identification submodule is configured to collect impact contact force data in real time based on sensors of the robot, identify the contact force data in a contact model through a particle swarm optimization algorithm, eliminate non-physical viscous force and optimize tangential velocity error, obtain contact parameters of the contact model, and the contact parameters comprise contact stiffness, contact damping, and a contact model coefficient.

[0153] The coupling contact model construction submodule is configured to construct a coupling contact model optimized based on a non-elastic impact contact model, obtain output normal impact force based on the contact parameters, describe the normal impact force based on a rate-dependent damping term, combine a modified Coulomb friction term to describe friction force, and obtain tangential friction force.

[0154] Optionally, the alternating impact disassembly operation robot dynamic characteristic modeling system further comprises:

[0155] The closed-loop checking module is configured to perform closed-loop checking on the alternating impact dynamics model through multi-source data fusion, collect postures, velocities, and contact forces of the robot in the impact process, extract velocity mutation characteristics after impact removal from the postures, velocities, and contact forces by polynomial fitting to obtain post-impact velocities, generate measured data, establish a multi-body dynamics simulation model, calculate a prediction error based on the measured data, analyze multi-space stability of the system through a joint Lyapunov index, and obtain an accuracy evaluation result of the alternating impact dynamics model.

[0156] It should be noted that the alternating impact disassembly operation robot dynamic characteristic modeling system provided by the embodiments of the present application can execute the alternating impact disassembly operation robot dynamic characteristic modeling method provided by any embodiment of the present application, and has the corresponding functions and beneficial effects of the execution method.

[0157] It has to be noted that, in the present document, relational terms are intended only to convey a possible relationship between elements or

[0158] The above description is merely that of the specific embodiments of the application and as such is not to be taken in a limiting sense. Various modifications and co nti n uations will be evident to those skilled in the art that do not depart from the spirit and scope of the application as defined by the appended claims. The specific embodiments presented, therefore, are not to be considered in a limiting sense, but are presented for purposes of illustration only, and numerous other embodiments are contemplated.

Claims

1. A method for modeling dynamic characteristics of a robot for disassembly of an alternating impact machine, characterized by, The method comprises the following steps: Real-time identification of robot joint side and motor side dynamic characteristics through an optimization algorithm, and updating the flexible joint model of the robot based on equivalent rigid inertia correction theory to obtain a joint dynamics equation fused with closed-loop torque control law, and output joint parameters of the robot, including joint side angle, acceleration and torque; Using a particle swarm optimization algorithm to identify the parameters of the contact model in combination with sensor data of the robot, and constructing a coupled contact model optimized based on a non-elastic impact contact model to obtain output normal impact force and tangential friction force; Based on the normal impact force, the tangential friction force and the joint parameters, a phase-by-phase modeling strategy is used to analyze the continuous contact phase and the impact phase, and an alternating impact dynamics model is constructed; Based on the alternating impact dynamics model, the non-smooth mechanics theory, impact dynamics and task constraints are combined to construct a coupled equation, and the impact dynamics equation is reconstructed through friction correction Jacobian matrix to obtain the post-impact velocity; Based on the post-impact velocity, an impact force control strategy is generated, and the working robot is controlled to perform an alternating impact contact task; The reconstructed impact dynamics equation is used to describe the velocity change characteristics of the robot during the impact process.

2. The method of claim 1, wherein, Real-time identification of robot joint side and motor side dynamic characteristics through an optimization algorithm, and updating the flexible joint model of the robot based on equivalent rigid inertia correction theory to obtain a joint dynamics equation fused with closed-loop torque control law, and output joint parameters of the robot, including: Analyzing the joint side and motor side dynamic characteristics during the impact process of the robot, and identifying the joint stiffness and motor inertia through frequency domain response data and an optimization algorithm in real time; Based on the joint stiffness and the motor inertia, a joint dynamics equation fused with closed-loop torque control law is constructed through equivalent rigid inertia correction theory based on the flexible joint model of the robot, and the joint parameters of the robot are output.

3. The method of claim 2, wherein, Based on the joint stiffness and the motor inertia, a joint dynamics equation fused with closed-loop torque control law is constructed through equivalent rigid inertia correction theory, and the joint parameters of the robot are output, including: With joint stiffness and motor inertia as input, based on the equivalent rigid inertia correction theory, according to the joint dynamics equation is constructed, and the angle , acceleration and torque on the joint side are obtained in real time; wherein, represents an externally applied joint torque, represents a motor-side angle, represents a joint-side angular velocity, represents a joint-space inertia, represents system Coriolis, centrifugal, and gravitational effects, represents a joint-side desired torque, represents a motor inertia.

4. The method of claim 1, wherein, Using a particle swarm optimization algorithm to identify the parameters of the contact model in combination with sensor data of the robot, and constructing a coupled contact model optimized based on a non-elastic impact contact model to obtain output normal impact force and tangential friction force, including: Real-time acquisition of impact contact force data by the sensors of the robot, and identification of the contact force data through a particle swarm optimization algorithm in the contact model to eliminate non-physical viscous force and optimize tangential velocity error, so as to obtain contact parameters of the contact model, including contact stiffness, contact damping and contact model coefficients; Constructing a coupled contact model optimized based on a non-elastic impact contact model, and obtaining output normal impact force based on the contact parameters; Describing the normal impact force based on a rate-dependent damping term, and obtaining tangential friction force by combining a modified Coulomb friction term.

5. The method of claim 4, wherein, Constructing a coupled contact model optimized based on a non-elastic impact contact model, and obtaining output normal impact force based on the contact parameters, including: Based on contact stiffness , contact damping , and contact model coefficients , according to , a coupled contact model optimized based on a non-elastic impact contact model is constructed to obtain an output normal impact force ; wherein, represents the contact deformation amount, represents the contact deformation rate.

6. The method of claim 5, wherein, The normal impact force is described based on a rate-dependent damping term, the friction force is described by combining a modified Coulomb friction term, and the tangential friction force is obtained, including: With normal impact force as input, according to , the tangential friction force is obtained; wherein, represents the contact deformation, represents the contact deformation rate, represents the friction coefficient, represents the impact contact and the resultant velocity of the directions, represents the vector module solution.

7. The method of claim 3, wherein, Based on the normal impact force, the tangential friction force and the joint parameters, an alternating impact dynamics model is constructed by using a phase separation modeling strategy to couple and analyze continuous contact phases and impact phases, including: Based on normal impact force , the mapping relationship between normal impact force and joint torque is obtained by coupling continuous contact force through flexible joint dynamics with the strategy of phase separation modeling ; Based on tangential friction force , the mapping relationship between the tangential friction force and joint torque is obtained by solving the velocity jump with the impulsive momentum equation based on the phase separation modeling strategy ; Based on the mapping relationship and the mapping relationship , according to , constructing an alternating impact dynamics model; wherein, , denotes the joint velocity after impact contact, denotes the joint velocity before impact contact, denotes the normal Jacobian matrix, denotes the tangential Jacobian matrix.

8. The method of claim 7, wherein, On the basis of the alternating impact dynamics model, a coupling equation is constructed by combining non-smooth mechanics theory and impact dynamics, task constraints, and the impact dynamics equation is reconstructed by modifying the Jacobian matrix through friction, and the post-impact velocity is obtained, including: On the basis of the alternating impact dynamics model, a multi-space dynamic coupling framework is constructed by combining non-smooth mechanics theory and impact dynamics; Based on the multi-space dynamic coupling framework and the Jacobian mapping of task constraints, according to , the coupling equations of joints, contacts and task space are constructed; Correcting a jacobian matrix by friction , updating coupling equations according to , reconstructing impact dynamics equations to obtain post-impact velocity ; wherein denotes the identity matrix.

9. The method of claim 1, wherein, After obtaining the post-impact velocity, it further includes: The alternating impact dynamics model is closed-loop checked through multi-source data fusion, and the posture, velocity and contact force of the robot during the impact process are collected; The post-impact velocity is obtained by using polynomial fitting to extract the velocity mutation characteristics after the impact is removed from the posture, velocity and contact force, and the measured data is generated; A multi-body dynamics simulation model is established, the prediction error is calculated based on the measured data, the multi-space stability of the system is analyzed through joint Lyapunov index, and the accuracy evaluation result of the alternating impact dynamics model is obtained.

10. An alternating impact disassembly operation robot dynamic characteristic modeling system, characterized by, Including: The joint dynamics equation construction module is used to identify the dynamic characteristics of the robot joint side and the motor side in real time through an optimization algorithm, and update the flexible joint model of the robot based on the equivalent rigid inertia correction theory, obtain the joint dynamics equation fused with the closed-loop torque control law, and obtain the output joint parameters of the robot, including the angle, acceleration and torque of the joint side; The coupling contact model construction module is used to identify the parameters of the contact model by combining the sensor data of the robot using a particle swarm optimization algorithm, and construct a coupling contact model optimized based on a non-elastic impact contact model, and obtain the output normal impact force and tangential friction force; The alternating impact dynamics model construction module is used to construct an alternating impact dynamics model by coupling and analyzing continuous contact phases and impact phases based on the normal impact force, the tangential friction force and the joint parameters, using a phase separation modeling strategy; The impact dynamics equation reconstruction module is used to construct a coupling equation by combining non-smooth mechanics theory and impact dynamics, task constraints, on the basis of the alternating impact dynamics model, and reconstruct the impact dynamics equation by modifying the Jacobian matrix through friction, and obtain the post-impact velocity; The impact force control module is used to generate an impact force control strategy based on the post-impact velocity, and control the work robot to perform an alternating impact contact task. The reconstructed impact dynamics equation is used to describe the velocity change characteristics of the robot during the impact process.

Images