CN-121979111-A - Truss robot previewing control method based on digital twin

CN121979111ACN 121979111 ACN121979111 ACN 121979111ACN-121979111-A

Abstract

The invention discloses a digital twin-based truss robot previewing control method which comprises the following steps of S1, constructing a truss robot digital twin system, constructing a truss robot virtual scene and constraint, synchronizing data and carrying out two-way communication, configuring a safety rule base and a data recording/playback interface at a twin end, S2, previewing an execution task, generating a reference track and a constraint set, automatically checking a previewing process according to the safety rule base to obtain a previewing passing reference track and the constraint set, S3, carrying out parameterization prediction control and real-time tracking on the previewing passing reference track and the constraint set, and controlling track tracking errors within a tolerance range. According to the invention, a digital twin system corresponding to the physical world is constructed, the model prediction control parameterized by the Laguerre function is introduced to realize low-dimensional quick solution, high-precision, reproducible and high-reliability previewing control of the truss robot is realized, and the field debugging and operation and maintenance costs are remarkably reduced.

Inventors

LIU MANLU
ZHENG YUANXIN
LI JINHAO
CHEN ZHILIANG
GUI LIPING
LIU HONGWEI
LIU YU
ZHOU RUI
LIU XIN
LONG RUI
HOU LINGFENG

Assignees

西南科技大学

Dates

Publication Date: 20260505
Application Date: 20251212

Claims (8)

1. The digital twin-based truss robot previewing control method is characterized by comprising the following steps of: S1, constructing a digital twin system of a truss robot, wherein the digital twin system comprises the steps of constructing a virtual scene and constraint of the truss robot, synchronizing data, bidirectionally communicating the virtual scene and constraint, configuring a safety rule base at a twin end and configuring a data recording/playback interface; S2, previewing the task to be executed, generating a reference track and a constraint set, and automatically checking the previewing process according to a safety rule base to obtain a previewed reference track and a constraint set; S3, performing parameterized predictive control and real-time tracking on the reference track and the constraint set which are previewed, and controlling the track tracking error to be within a tolerance range.
2. The digital twin truss robot previewing control method according to claim 1, wherein in S1, the method for constructing a virtual scene and constraints of the truss robot comprises: Geometric modeling and light weight importing, namely geometric modeling is carried out on a truss body, a mobile platform, an operation object and an end effector, 3ds Max is adopted for format conversion and part light weight, and a coordinate axis is converted and then imported into Unity 3D by fbx; Establishing a father-son hierarchical relationship and consistency check, namely establishing a father-son hierarchical structure among joints in the Hierarchy of the Unity 3D to ensure that a virtual model and a physical robot keep consistent in behavior; The coordinate system and joint parameter configuration, namely, a left-hand coordinate system is adopted by Unity 3D, and on the premise of ensuring that a Z axis is a Rotation axis, positions and rotations of all joints on a Y axis are zeroed to meet the kinematic coordinate condition, and kinematic model parameters are configured on the basis; The constraint model is configured by setting a multi-level collision bounding box based on physical attribute dimensions to realize collision detection and basic safety constraint of a twin body, and combining the movement ranges of all joints in a DH parameter table to form joint limiting constraint.
3. The truss robot previewing control method based on digital twin according to claim 2, wherein in S1, the data synchronization and the bidirectional communication comprise configuring real-time data interfaces among the physical robot, the sensor and the virtual twin, supporting the real-time reporting of the robot state and the safe issuing of control instructions, and integrating the history data record and playback.
4. The truss robot previewing control method based on digital twin according to claim 1, wherein in S2, the method for previewing the task to be executed is specifically: The method comprises the steps of arranging passing path points according to task demands, establishing a kinematic chain with father-child layers and DH parameters of a joint of a mechanical arm on the basis of XYZ three-axis coordinates according to a truss to form accessibility and travel constraint to obtain a constraint set, executing path search and track time parameterization in a virtual environment, performing whole-course animation previewing on joint actions of a base and the mechanical arm, recording evaluation values of a joint tail end curve, a path length, execution time, a minimum gap and smoothness according to a fixed period, and generating a reference track comprising joint position, speed, acceleration and tail end precision; The reference track and the constraint set are used for automatically checking the problems of collision, limit, singular, travel and communication delay according to the safety rule base.
5. The digital twin based truss robot previewing control method according to claim 1, wherein S3 comprises the following sub-steps: s31, unifying data types of entities and twins, and finishing initial pose and scale alignment, constructing a playable historical database according to periodic sampling, control quantity and actual errors, and constructing a Laguerre function parameterized model prediction controller; S32, inputting a reference track and a constraint set into a Laguerre function parameterized model prediction controller, performing rolling optimization solution through parameterized prediction control, and explicitly processing constraints of travel, speed and acceleration; S33, judging whether the track tracking error is within the tolerance range, if so, solidifying the current optimal control quantity to be a new running base line to complete entity track tracking, and if not, starting a playback-retest-reengineer flow to update the reference track and the constraint set, and returning to S32.
6. The digital twin truss robot previewing control method according to claim 5, wherein in S31, the method for constructing the model predictive controller parameterized by the lager function specifically comprises: a1, initializing a target and establishing a historical database: The data caliber of the entity end and the twin end is unified, namely, the shaft/joint state in a PLC register is read and analyzed into standard floating point quantity according to Modbus TCP, the standard floating point quantity is mapped to a unified coordinate and unit system, a JSON structured message body is adopted on a twin decision side to describe 'state/control' data, a historical database is established on the basis, original sampling, reference/control quantity and error information are continuously written according to a control period, and time information is added to support track visualization and playback, scene restoration and comparison analysis; A2, synchronizing virtual and real states: a21, aligning the initial postures of the virtual model and the physical robot through position calibration, measuring distances between the XYZ three-axis limit positions of the truss and a plurality of preset alignment points, and recording pose data of the virtual end and the physical end; A22, entering a bidirectional control synchronous test of the entity end and the twin end, recording the coordinate values of the XYZ triaxial positions from the entity end to the datum point at the alignment point, comparing with the twin end reference time by time, counting error samples, judging whether the average error obtained by the error samples exceeds an error threshold, if not, completing the construction of the Laguerre function parameterized model predictive controller, if so, changing the position calibration parameters, and returning to A21.
7. The digital twin truss robot previewing control method according to claim 6, wherein in S32, the model predicts the objective function of the controller The expression of (2) is specifically: In the formula, In order to control the input weight matrix, For the state weight matrix, C is the linear term coefficient vector in the objective function, For a mapping matrix consisting of discrete laguerre orthogonal functions, For the prediction matrix, U is the control quantity, In order for the upper limit of the actuator to be set, In order for the lower limit of the actuator to be set, For the lower limit of the coefficient of the laguerre, For the upper limit of the Laguerre coefficient, T is the transposed symbol, In order for the coefficient of the laguerre, As a scalar relaxation factor, the degree of the relaxation factor, As the weight coefficient of the weight coefficient, ; In the formula, As the order of the basis function, In order to predict the time domain of the signal, In order to control the time domain of the signal, In order to be a state response constraint, In order to control the constraint of the variables, For the purpose of terminal state constraints, As a state matrix for an incremental system, Is a control input matrix of an incremental system, I is an identity matrix, In the form of a state propagation matrix, Is an incremental state vector at time k, Transpose the vector for the lager function at time k, , Is that The Laguerre basis function row vector corresponding to the 1 st prediction step obtained by recursion in (1), Is that The Laguerre basis function row vector corresponding to the 2 nd prediction step obtained by recursion in (2), Is that A Laguerre basis function row vector corresponding to the 3 rd prediction step obtained by recursion; each sampling period of the model predictive controller takes a previewed reference track and constraint set as input, and solves the order of the basis function Is optimized for (a) Sequence and reverse conversion to optimal control increment As an output; In the formula, For the X-axis number vector calculated at time k, For the Y-axis number vector calculated at time k, The Z-axis number vector calculated for the k moment, For the optimal control increment of the X axis at the moment k, For the Y-axis optimal control increment at time k, And optimally controlling the increment for the Z axis at the moment k.
8. The truss robot previewing control method based on digital twin according to claim 5, wherein in S33, the playback-retest-reenacting flow is specifically: and generating a tracking track and track tracking error curve according to the current optimal control quantity, rechecking the scale and the zero position by combining a position calibration experiment, and re-previewing and checking at the virtual end according to the reference track and the constraint set to generate an updated reference track and constraint set.

Description

Truss robot previewing control method based on digital twin Technical Field The invention belongs to the technical field of truss robot control, and particularly relates to a digital twin-based truss robot previewing control method. Background The truss robot has the advantages of large working space, strong load capacity, high movement precision and the like, and is widely applied to links such as workshop assembly, transportation, storage, inspection and the like. For related applications of truss robots, the existing research is mainly dependent on platforms such as MATLAB/Simulink, ADAMS, gazebo, coppelia Sim and the like to conduct offline simulation previewing, and the method can accelerate algorithm iteration, but lacks real-time bidirectional mapping and running state feedback with a real system, so that tracks and control parameters obtained by simulation are difficult to directly reproduce, and a large amount of field debugging and manual intervention are still needed. In recent years, digital twin systems oriented to truss robots are beginning to be used for virtual-real mapping, visual monitoring and task planning, and some researches introduce high-fidelity modeling and coordinated tracking control, and try to use some control algorithms to reduce control cycle delay and improve tracking accuracy. However, most of the existing schemes still stay in the links of visualization, offline checking and local function realization, have the problems of previewing deviation and communication time delay caused by inaccurate models or uncertain parameters, lack of online error compensation and safety blocking for high-precision operation, and the like, and are difficult to meet the requirements of workshop scenes on high precision, reproducibility and implementation. Disclosure of Invention Aiming at the defects in the prior art, the digital twin-based truss robot previewing control method provided by the invention constructs a digital twin system corresponding to the physical world and establishes a virtual-real bidirectional mapping mechanism with low time delay, performs task previewing in advance in the twin domain, introduces Laguerre function parameterized model prediction control to realize low-dimensional quick solution so as to realize on-line error compensation and look-ahead control, performs self-adaptive updating on a reference track through real-time feedback, realizes high-precision, reproducible and high-reliability previewing control of the truss robot in the scenes of assembly, inspection and the like of a manufacturing workshop, obviously reduces the field debugging and operation cost, and solves the problems that the existing scheme lacks a bidirectional mapping and on-line synchronization mechanism and is difficult to meet the requirements of workshop scenes on high precision, reproducible and implementation. In order to achieve the aim of the invention, the technical scheme adopted by the invention is that the digital twin-based truss robot previewing control method comprises the following steps: S1, constructing a digital twin system of a truss robot, wherein the digital twin system comprises the steps of constructing a virtual scene and constraint of the truss robot, synchronizing data, bidirectionally communicating the virtual scene and constraint, configuring a safety rule base at a twin end and configuring a data recording/playback interface; S2, previewing the task to be executed, generating a reference track and a constraint set, and automatically checking the previewing process according to a safety rule base to obtain a previewed reference track and a constraint set; S3, performing parameterized predictive control and real-time tracking on the reference track and the constraint set which are previewed, and controlling the track tracking error to be within a tolerance range. Further, in S1, the method for constructing the virtual scene and the constraint of the truss robot comprises the following steps: Geometric modeling and light weight importing, namely geometric modeling is carried out on a truss body, a mobile platform, an operation object and an end effector, 3ds Max is adopted for format conversion and part light weight, and a coordinate axis is converted and then imported into Unity 3D by fbx; Establishing a father-son hierarchical relationship and consistency check, namely establishing a father-son hierarchical structure among joints in the Hierarchy of the Unity 3D to ensure that a virtual model and a physical robot keep consistent in behavior; The coordinate system and joint parameter configuration, namely, a left-hand coordinate system is adopted by Unity 3D, and on the premise of ensuring that a Z axis is a Rotation axis, positions and rotations of all joints on a Y axis are zeroed to meet the kinematic coordinate condition, and kinematic model parameters are configured on the basis; The constraint model is configured by setting a multi-level collision bounding b