Singularity Prediction for Parallel Robots for Improvement of Sensor-Integrated Assembly

The frequent occurrences of singularities inside a parallel robot's workspace and in certain cases of sensor integration non-deterministic end-effector trajectories demand a powerful online singularity prediction and a control design that is capable of displacing motion algorithms. The theory of a power-inspired index of closeness to a singularity pose for parallel kinematic manipulators is illumined and presented in this paper, supported by comparison to Grassmann geometry and experimental results achieved on a HEXA robot. An experimentally validated strategy for jerk bounded braking the end-effector while approaching a singularity and an algorithm for departing it again are presented as well.