Manipulator motion planning method based on bias redefined neural network

• Main Authors: ZHANG ZHIJUN, LI ZEYANG, ZHENG LUNAN, CUI YAOWEI, SU SHUOMIAO
• Format: Patent
• Language: Chinese; English
• Published: 24.11.2020
• Subjects: CHAMBERS PROVIDED WITH MANIPULATION DEVICES; HAND TOOLS; MANIPULATORS; PERFORMING OPERATIONS; PORTABLE POWER-DRIVEN TOOLS; TRANSPORTING

The invention discloses a manipulator motion planning method based on a bias redefined neural network. The manipulator motion planning method based on the bias redefined neural network comprises the steps of establishing a kinematics equation of a redundant manipulator according to a model and parameters of an actual manipulator; transforming the kinematics equation of the redundant manipulator into a kinematics equation of a velocity layer, and describing the kinematics equation of the velocity layer as a smooth time-varying linear equation; acquiring a joint angle of the manipulator and endtrack information, constructing a time-varying parameter matrix according to the parameters of the manipulator, the joint angle of the manipulator and the end track information, and acquiring a corresponding time derivative; designing a bias function according to the smooth time-varying linear equation; designing a time-varying linear equation solver for a redundant robot arm by redefining a recurrent neural network with