Adaptive Biological Neural Network Control and Virtual Realization for Engineering Manipulator.
Journal:
Computational intelligence and neuroscience
PMID:
36072724
Abstract
By analyzing the feasibility of the digital twin technology in the assembly of construction machinery, the assembly process of the construction manipulator in the engineering environment is discussed. According to the application criteria and modeling requirements of digital twin, the overall framework of digital twin engineering manipulator assembly modeling and simulation is constructed from three aspects: model layer, data layer, and application layer. According to the operation task characteristics of space engineering manipulator, the feasibility of the control method based on joint angular velocity is analyzed, and the task environment of space engineering manipulator based on Markov model is defined. Aiming at the application of the algorithm in the control task of the space engineering manipulator, a reward function with the addition of the angular velocity soft bound term is designed, which improves the strategy optimization process of the algorithm and obtains a better control effect of the engineering manipulator. The motion trajectory of the end of the engineering manipulator is directly given on the simulation platform, and the expected motion of each joint of the engineering manipulator is calculated through the kinematics of the engineering manipulator. It can be seen from the simulation results that the controllers designed in this study can achieve ideal control effects. With the help of Baxter robot platform, the control algorithm designed in this study is applied to the actual engineering manipulator control, and the effectiveness of the control algorithm is further proved by the actual control effect.