A Repeatable Motion Scheme for Kinematic Control of Redundant Manipulators.

Area of Science:

• Robotics
• Control Systems
• Artificial Intelligence

Background:

• Achieving closed trajectory motion for redundant manipulators requires joints to return to initial positions.
• Existing repeatable motion schemes often overlook the practical difficulty of pre-positioning robot arm joints accurately.
• Effective kinematic modeling is crucial for precise and repeatable robot movements.

Purpose of the Study:

• To develop a novel optimal programming index for redundant manipulator motion planning.
• To formulate a repeatable motion scheme that addresses the challenges of joint pre-positioning.
• To enhance the accuracy and timeliness of robot trajectory execution.

Main Methods:

• Designed a novel optimal programming index using a recurrent neural network.
• Formulated a repeatable motion scheme incorporating kinematic equation constraints.
• Applied the Lagrange multiplier theorem to convert the scheme into a time-varying linear equation.
• Developed a finite-time neural network solver for the motion scheme.

Main Results:

• The proposed scheme demonstrates accuracy and timeliness in simulation for diverse trajectories.
• The recurrent neural network effectively optimizes the kinematic model for repeatable motion.
• The finite-time neural network solver efficiently solves the complex motion planning problem.

Conclusions:

• The novel kinematic scheme offers an effective solution for closed trajectory motion planning in redundant manipulators.
• The integration of recurrent neural networks and Lagrange multipliers provides a robust and efficient approach.
• The proposed method significantly improves the repeatability and timeliness of robot arm movements.