Toward coordinated planning and hierarchical optimization control for highly redundant mobile manipulator

Area of Science:

• Robotics
• Control Systems
• Artificial Intelligence

Background:

• Mobile manipulators require advanced control for complex indoor environments.
• Traditional optimization methods struggle with joint limits and constraints in redundant systems.

Purpose of the Study:

• To develop a constraint planning and optimization control scheme for highly redundant mobile manipulators.
• To improve the handling of joint limits and equality constraints in manipulator control.
• To enable effective path planning and tracking control in complex indoor environments.

Main Methods:

• Introduced infinity norm and slack variables into the optimization algorithm.
• Expressed the tracking control problem as a quadratic programming (QP) problem using derived kinematic equations.
• Proposed a two-timescale recurrent neural networks optimization scheme.
• Integrated the BI²RRT* path-planning algorithm for complex environments.

Main Results:

• Successfully tested the optimization scheme on a 9 Degrees of Freedom (DOFs) nonholonomic mobile-based manipulator.
• Demonstrated effective path planning and tracking control in simulated complex indoor environments.
• Validated the approach using predefined and generated paths in the Neurorobotics Platform (NRP).

Conclusions:

• The proposed constraint planning and optimization control scheme effectively manages joint limits and constraints.
• The two-timescale recurrent neural networks optimization scheme provides a robust solution for mobile manipulator control.
• The BI²RRT* algorithm enhances path planning capabilities in cluttered indoor spaces.