Stability regime and comparison with quasi-static estimators

Determine the stability regime of the Bayesian continuum robot dynamics factor-graph optimizer and compare its stability and estimation behavior with quasi-static estimators.

Background

The proposed factor-graph optimizer was stable on the trajectories reported in the paper but failed to converge on some aggressive trajectories, which the authors attributed to poor conditioning. The paper addresses this issue through stress parameterization and an L-stable temporal discretization, but does not characterize the conditions under which the optimizer is stable or fails.

The authors also identify a need to compare the dynamic estimator systematically with quasi-static estimators. This comparison is relevant because quasi-static models may be adequate for slow, lightweight robots, whereas inertial modeling is important when inertial forces are comparable to external loads.