An adaptive link position tracking controller for rigid-link flexible-joint robots without velocity measurements

Area of Science:

• Robotics
• Control Systems Engineering
• Mechanical Engineering

Background:

• Rigid-link flexible-joint (RLFJ) robots present complex control challenges due to inherent uncertainties.
• Accurate state estimation, particularly velocities, is often difficult and costly in RLFJ systems.
• Existing controllers may require full state information, limiting practical applications.

Purpose of the Study:

• To develop an adaptive partial state feedback controller for RLFJ robots.
• To address parametric uncertainty in the mechanical system.
• To achieve robust and accurate robot control using limited sensor measurements.

Main Methods:

• Utilized a set of filters to estimate unmeasurable velocities (link and actuator).
• Employed an adaptive integrator backstepping procedure for controller design.
• Developed a torque input controller based on filtered states.

Main Results:

• The proposed controller compensates for parametric uncertainty in RLFJ robots.
• Achieved semiglobal asymptotic link position tracking.
• Ensured all closed-loop signals remain bounded, demonstrating stability.

Conclusions:

• The adaptive partial state feedback controller effectively manages RLFJ robot dynamics.
• The controller's reliance on position measurements simplifies implementation.
• Simulation results validate the controller's performance for a two-link RLFJ robot.