Related Experiment Videos
A unified quadratic-programming-based dynamical system approach to joint torque optimization of physically
Summary
This study presents a new quadratic programming (QP) solver for optimizing redundant robot arm torques under physical constraints. The efficient QP-based dynamical system solver avoids matrix inversion and enhances control accuracy.
Area of Science:
- Robotics
- Control Systems
- Optimization
Background:
- Redundant manipulators require complex joint torque optimization to respect physical limitations.
- Existing redundancy resolution schemes often involve computationally intensive methods.
Purpose of the Study:
- To formulate joint torque optimization for redundant manipulators as a quadratic programming (QP) problem.
- To introduce an efficient online solver for the QP problem using primal-dual dynamics.
- To demonstrate the solver's effectiveness and advantages over previous methods.
Main Methods:
- Formulating velocity-level and acceleration-level redundancy resolution as QP problems with constraints.
- Developing a primal-dual dynamical system solver based on linear variational inequalities.
- Simulating the proposed QP-based dynamical system on a PUMA560 robot arm.
Main Results:
- The proposed QP solver exhibits simple piecewise-linear dynamics.
- The solver eliminates the need for real-time matrix inversion.
- Joint acceleration information is provided for torque control in velocity-level schemes.
- Simulations confirm the efficiency and effectiveness of the approach.
Conclusions:
- The QP-based dynamical system offers an efficient and effective method for redundant manipulator joint torque optimization.
- The solver's characteristics make it suitable for real-time online control applications.
- This approach enhances manipulator control by providing crucial joint acceleration data.