Distributed Neural-Network-Based Cooperation Control for Teleoperation of Multiple Mobile Manipulators Under...

Area of Science:

• Robotics
• Control Systems Engineering
• Artificial Intelligence

Background:

• Teleoperation systems with multiple slave robots face challenges like communication delays and bandwidth limitations.
• Cooperative control is crucial for synchronized manipulation and formation tasks in these systems.
• Existing methods often struggle with the complexities of distributed control and dynamic uncertainties.

Purpose of the Study:

• To design a fully distributed cooperative control strategy for sampled-data teleoperation systems with multiple slave mobile manipulators.
• To ensure task-space position synchronization between the master and the grasped object, maintaining a fixed formation.
• To address communication bandwidth limitations and time-varying delays inherent in discrete-time transmissions.

Main Methods:

• A distributed control strategy employing neural-network-based controllers for task-space synchronization and null-space formation.
• Utilization of Radial Basis Function (RBF) neural networks for adaptive compensation of dynamical uncertainties.
• Application of the Round-Robin (RR) scheduling protocol to manage data transmission from multiple slaves to the master.
• Analysis of system stability and synchronization/formation features, establishing relationships between control parameters, time delays, and sampling intervals.

Main Results:

• The proposed control strategy effectively achieves task-space position synchronization and fixed formation control for the single-master-multiple-slaves (SMMS) teleoperation system.
• Adaptive RBF neural networks successfully compensated for dynamical uncertainties, enhancing system robustness.
• A clear relationship was established between control parameters, time delay bounds, and the maximum sampling interval, providing design guidelines.
• Experimental validation confirmed the effectiveness and practical applicability of the developed control algorithm.

Conclusions:

• The developed distributed cooperative control strategy is effective for sampled-data teleoperation systems with multiple mobile manipulators.
• The use of adaptive RBF neural networks provides a robust solution for handling uncertainties in complex teleoperation tasks.
• The findings offer valuable insights into the design and stability analysis of multi-robot teleoperation systems operating under communication constraints.