Event

State estimation pipeline proposed for hybrid rigid-flexible space manipulator

1 reports · 1 sources

Lorenzo Ticozzi and co-authors propose a state reconstruction pipeline for a hybrid rigid-flexible space manipulator whose distal link is a tendon-driven continuum robot. The method formulates the estimation problem as maximum a posteriori inference conditioned on joint encoders and visual measurements, and transcribes it into a factor graph, fusing heterogeneous measurements with a rigid-flexible kinematic model, while employing the geometrically variable strain (GVS) method to obtain a minimal-order configuration parameterization. The implementation is validated only in simulation: state reconstruction accuracy and computational performance are examined on closed-loop trajectories, and position measurements of the continuum link come from a simulated vision pipeline using synthetic photorealistic images.