This paper presents a solution scheme for forward kinematics of a general fully parallel manipulator that guarantees a unique solution with only three redundant sensors. The redundant sensors were designed to minimize engineering difficulties in the realization, whereas an optimal sensor location was proposed to achieve a numerical efficiency and accuracy significantly better than existing solutions. As a result, the present approach is insensitive to misalignment of sensor location and measurement errors. Due to these merits, the present approach can handle joint misalignment due to machining error and assembly when it is applied to the task of self-calibration, which requires extreme precision. It is also suitable for applications in which real-time computation is needed.

No abstract

No abstract

No abstract

No abstract

No abstract

No abstract

In this paper, two inverse-kinematic solutions with singularity robustness for a fully parallel robot manipulator are proposed. First, two inverse-kinematic solutions are presented based on the damped leastsquares method, where the first solution (OSVC) is derived using a velocity constraint, and the other (OSFC) uses a force constraint. It can be shown that both solutions are special cases of a general damped least-squares method (GIKS). Next, under the assumption that direct position measurement is available, a feedback-correction term is introduced into the general damped least-squares method to eliminate positioning error in resolved-rate control. It can be shown that both solutions are stable with bounded errors, but each has its own relative merits and shortcomings. Illustrative examples are also given to show their effectiveness in practical applications.