Experimental Forward Kinematics of a Cable-Driven Parallel Robot with Pose Estimation Covariance Bounds
The purpose of my work in the fall (2021) was to expand on a previous algorithm created to test the forward kinematics methods on experimental data of a cable-driven parallel robot (CDPR). The algorithm was developed and tested with two distinct methods and stopping criteria in simulation and proved to be accurate when determining the covariance bounds of the pose estimate errors of a CDPR's payload. This work uses experimental data of the pose of a payload and tests the covariance bounds on two parameters: the Euler Angle Sequence and the Quaternion. It was found that the covariance bounds accurately encapsulated the pose of the payload for both parameters. This progress and the work completed during the summer were published as a journal paper to IEEE in November. Current work this spring considers calibration errors in the system on top of the previously studied cable length measurement noise. The updated forward kinematics formulation is currently producing inconsistent results on improving the pose estimates and is under revision before testing the accuracy over multiple runs.