Abstract
This paper presents a control method for managing both the position of a robot manipulator and its contact force. The proposed approach incorporates a virtual impedance model to correct errors between reference signals and control values. Evaluated through state-space derivations and simulations, the method demonstrates that a robot arm can effectively follow target position and force values. The results confirm the effectiveness of integrating virtual impedance into the control framework for improved stability and performance.
Introduction
In environments where robots interact closely with humans, such as medical care and nursing, "force-position" control is paramount. A robot must reach a specific spatial target while simultaneously regulating the force it exerts to avoid causing injury or damage. Traditional hybrid control methods often treat position and force as decoupled tasks, which can lead to operational interference. This research addresses these challenges by employing Linear Quadratic Integral (LQI) control enhanced with a "virtual impedance" model, allowing the robot to dynamically react to environmental stiffness and friction.
Methodology
The researchers developed a mathematical model for a 2-link robotic arm interacting with a vertical surface. The core of the approach lies in Virtual Impedance Modeling, where the contact point is represented by virtual springs ($K_{ex}, K_{ey}$) and dampers ($C_{ex}, C_{ey}$). This creates a digital buffer that the controller uses to calculate necessary force adjustments based on positional errors.
The system's equations of motion incorporate the Jacobian matrix to map joint torques to Cartesian forces, accounting for both static and dynamic friction. The LQI Control Design utilizes an integral term to eliminate steady-state errors, ensuring the arm converges precisely to its target.
Key Insights and Results
The effectiveness of the proposed method was evaluated through numerical simulations in MATLAB.
Stability and Eigenvalue Analysis
A critical insight from the study is the effect of virtual impedance on system stability. By increasing the virtual impedance, the system's eigenvalues shift further into the left-half of the complex plane. This shift indicates a more stable system with a faster transient response, as confirmed by the eigenvalue comparison in the study.
Performance Comparison
Simulations demonstrated significant improvements in position tracking ($P_x$). As shown in Figure 2, the arm reaches its target position of 0.3m much faster and with considerably less oscillation when virtual impedance is considered compared to when it is ignored.
While position control saw marked improvement, force tracking ($F_y$) presented a more complex result. Both methods successfully converged to the 2N target, but the version without virtual impedance reached the target force faster in this specific simulation setup. This suggests that while stability and position accuracy are enhanced, the force response timing is sensitive to parameter tuning.
Conclusion and Future Work
The integration of virtual impedance into an LQI framework provides a robust solution for simultaneous position and force control in robotic arms. However, the study highlights the challenge of parameter tuning; the values for virtual springs, dampers, and LQI weight matrices currently require manual adjustment. Future research will focus on developing automated methods for calculating these design parameters to optimize the balance between positional precision and force response speed.
Summary updated