CN-122008208-A - Double-arm robot cooperative control method based on cooperative index and quadratic programming

CN122008208ACN 122008208 ACN122008208 ACN 122008208ACN-122008208-A

Abstract

The invention discloses a cooperative control method of a double-arm robot based on cooperative indexes and quadratic programming, which belongs to the technical field of robot control and comprises the steps of constructing a cooperative index model, fusing a track tracking error, a double-arm synchronous error and a tail end contact force error by the cooperative index model, introducing a dynamic weight factor to continuously switch control modes, generating a flexible offset track by adopting admittance control according to the cooperative index model, updating the cooperative index model by taking the flexible offset track as a track correction item, constructing a quadratic programming problem comprising joint moment, tail end speed and contact force constraint with the aim of minimizing the cooperative indexes, solving the quadratic programming problem in real time in each control period, and obtaining an optimal joint control moment and outputting the optimal joint control moment to the double-arm robot. The invention realizes flexible multi-task switching and accurate control.

Inventors

LI SHIQI
Gu Leyuan
LI XIAO
HU YUFENG
WANG LEI
XU ZHIYUAN
CHE ZHENGPING
XIONG YOUJUN
TANG JIAN

Assignees

华中科技大学
北京人形机器人创新中心有限公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1. A double-arm robot cooperative control method based on cooperative index and quadratic programming is characterized by comprising the following steps: Constructing a collaborative index model, wherein the collaborative index model fuses a track tracking error, a double-arm synchronous error and a tail end contact force error, and introduces a dynamic weight factor to continuously switch control modes; Generating a compliant offset track by adopting admittance control according to the collaborative index model, and updating the collaborative index model by taking the compliant offset track as a track correction term; Constructing a quadratic programming problem containing joint moment, terminal speed and contact force constraint by taking the minimum cooperative index as a target; And solving the quadratic programming problem in real time in each control period, obtaining the optimal joint control moment and outputting the optimal joint control moment to the double-arm robot.
2. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process for constructing the collaborative index model comprises the following steps: Acquiring respective track tracking errors of the two arms, movement synchronization errors between the two arms and terminal contact force errors at the current moment; Dynamically adjusting the cooperative weight factor and the force weight factor according to the task stage; and linearly fusing the track tracking error, the inter-arm motion synchronization error and the tail end contact force error with the cooperative weight factor and the force weight factor to obtain a cooperative index.
3. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of dynamically adjusting the collaborative weight factor and the force weight factor comprises: If the cooperative weight factor and the force weight factor are both 0, setting a track tracking mode; if the cooperative weight factor is greater than 0 and the force weight factor is 0, setting to a double-arm synchronous motion mode; If the cooperative weight factor and the force weight factor are both greater than 0, setting the cooperative weight factor and the force weight factor as a force-bit mixed control mode; and if the cooperative weight factor is 0 and the force weight factor is greater than 0, setting to a force priority control mode.
4. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of generating compliant deflection tracks using admittance control includes: Acquiring a force error between the actual contact force and the expected contact force of the tail end of the robot; inputting the force error into an admittance model, calculating a desired acceleration of the mass-damping-spring system; and carrying out discrete integration on the expected acceleration to obtain a compliant offset track.
5. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of updating the collaborative index model by taking the compliant offset trajectory as a trajectory correction term comprises the following steps: Superposing the compliant offset track to an original expected track to obtain a corrected reference track; Recalculating a track tracking error according to the corrected reference track; and replacing the original track tracking error with the track tracking error, and reconstructing the cooperation index.
6. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of constructing a quadratic programming problem containing joint moment, tip speed and contact force constraints, targeting minimizing the collaboration index, includes: Establishing a linear mapping relation between joint moment and terminal position increment; Expressing the cooperation index as a linear function of the joint moment; and constructing a quadratic objective function taking joint moment as an optimization variable.
7. The method of claim 6, wherein the step of providing the first layer comprises, The process for constructing the quadratic objective function taking the joint moment as an optimization variable comprises the following steps: calculating a quadratic term coefficient and a first term coefficient of the objective function according to the mapping relation of the cooperative index to the joint moment; And ignoring the constant term to obtain a standard quadratic programming form.
8. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of constructing a quadratic programming problem containing joint torque, tip speed, and contact force constraints further includes: The joint moment limit, the tip speed limit, and the tip contact force limit are written as linear inequalities, respectively, and then merged into a unified linear inequality constraint.
9. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of solving the quadratic programming problem in real time in each control period includes: in each control period, updating the objective function coefficient and the constraint condition according to the current robot state; Calling a solver to solve the standard quadratic programming problem on line to obtain an optimal joint moment; And outputting the optimal joint moment to each joint driver to realize moment control.
10. The method of claim 1, wherein the step of determining the position of the substrate comprises, The method further comprises the step of circularly executing the collaborative index model, admittance control and the quadratic programming problem in the same controller to form closed-loop multi-mode collaborative control.

Description

Double-arm robot cooperative control method based on cooperative index and quadratic programming Technical Field The invention belongs to the technical field of robot control, and particularly relates to a double-arm robot cooperative control method based on cooperative index and quadratic programming. Background Compared with a single-arm robot, the double-arm robot has the advantages of strong coordination operation capability, wide coverage of a working range, flexible system structure organization and the like, and is widely applied to complex tasks such as flexible assembly, cooperative transportation, man-machine interaction and the like. However, the high task adaptability of the double-arm robot obviously improves the difficulty of cooperative control, namely, on one hand, the requirement of precision of track tracking and synchronous movement between the double arms is met, and on the other hand, real-time and flexible response to external contact force change is also required. Specifically, at first, as the current control strategy cannot simultaneously consider multiple control targets such as track tracking, double-arm synchronization, end force feedback and the like, the modules lack unified modeling, so that conflict and switching obstacle exist between control tasks. Secondly, the switching of the control mode mostly depends on discrete heuristic judgment, and if a certain threshold is set, the mode switching is triggered if the threshold is exceeded. Therefore, a continuously adjustable and measurable unified index is lacked to drive the allocation of control targets in different task stages, so that the response of the system is unstable and the control behavior is unpredictable. Finally, although the traditional force-bit mixed control and impedance/admittance control have good flexibility, the parameter adjustment and task control targets are not related, and a coupling mechanism for track tracking control and double-arm motion synchronous control is lacked, so that the traditional force-bit mixed control and impedance/admittance control are difficult to integrate into a high-precision multi-target parallel task scene, and the practical usability in complex interactive operation is limited. Currently, some designs and researches have been made on a cooperative control method of a dual-arm robot by the related art. CN118081766a proposes a master-slave unified admittance control method of a double-arm robot for coordination tasks. The method comprises the steps of obtaining a terminal expected track by establishing an open-chain and closed-chain system kinematic model, decomposing the resultant force of an object, introducing a grabbing matrix and admittance control to design a master-slave unified admittance control algorithm, and establishing a cooperative control strategy aiming at non-coordinated, loose-coordinated and tight-coordinated tasks so as to keep the internal force of the system stable under man-machine cooperation and external force interference, so that the flexible operation of the double-arm robot is realized, and the method can be applied to scenes such as cooperative transportation. However, although a unified cooperative control framework is established for non-coordination, loose coordination and tight coordination tasks, a unified index is not established in the control process to comprehensively measure and coordinate multiple targets such as track tracking, motion synchronization and force feedback. In a practical complex task, it is difficult to intuitively evaluate and balance the relationship between the control targets. CN115625711B proposes a two-arm robot cooperative control method considering the tip force. Firstly, determining a motion constraint relation, then carrying out stress analysis and dynamics modeling through a Newton second law and a Lagrange method, then realizing position control by combining an interference observer and a hyperbolic tangent sliding mode controller, simultaneously obtaining expected contact force and an expected position through self-adaptive impedance control, and finally outputting joint moment after correction to complete a carrying task. The method utilizes the impedance model to adjust the end force response so as to complete the double-arm carrying task, has certain flexibility, is suitable for completing complex operation tasks in uncertain and rigid environments, but lacks unified modeling on multitasking targets (such as track, synchronization and force). Secondly, the control strategy takes the optimized tail end contact force as a main target, other control tasks cannot be flexibly adapted, and when the track tracking or double-arm synchronous motion task is completed, track tracking errors and synchronous motion errors are easy to occur, so that the task fails. CN118700151a proposes a two-arm robot cooperative control method, apparatus, electronic device and medium based on adaptive prediction. Firstly, motion data a