Adaptive Safety-Based Tracking Control for Uncertain Robotic Systems With Input-Output Constraints: A Neural...

Area of Science:

• Robotics
• Control Systems Engineering
• Artificial Intelligence

Background:

• Robotic systems often face challenges with uncertain dynamics, limited control torque, and joint position constraints.
• Ensuring safe operation while achieving precise trajectory tracking is crucial for practical robotic applications.

Purpose of the Study:

• To develop a novel control strategy for uncertain robotic systems that addresses input-output constraints and ensures safety.
• To enhance trajectory tracking performance in the presence of system uncertainties and physical limitations.

Main Methods:

• A neural network-based augmented high-order control barrier function (NN-AHoCBF) is proposed to estimate and compensate for system uncertainties.
• Adaptive bounds for neural network approximation errors and weights are incorporated into the high-order time derivative of the control barrier functions.
• Auxiliary systems are designed to adjust time-varying functions, relaxing control input constraints within the NN-AHoCBF framework.
• An adaptive safety-based tracking control method is formulated within a quadratic programming (QP) framework.

Main Results:

• The NN-AHoCBF method effectively estimates uncertainties and adapts to system dynamics.
• The proposed control strategy simultaneously satisfies input-output constraints, ensuring system safety.
• The controller demonstrates robust performance and accurate trajectory tracking capabilities.
• Simulations on a two-DOF robotic manipulator validate the controller's effectiveness.

Conclusions:

• The developed NN-AHoCBF provides an effective solution for trajectory tracking control of uncertain robotic systems with constraints.
• The adaptive safety-based approach enhances robustness and tracking accuracy while guaranteeing system safety.
• This method offers a promising direction for advanced robotic control applications facing complex operational conditions.