The abstract is read at the publisher; the summary is JClub's.
Feature-based readout adaptation in a simulated torque-driven planar two-link arm shows retained compensation after updates stop, with tracking error reduced by 40–97% versus pre-adaptation.
- Why it matters: It addresses how much an adaptive controller truly retains when the force field is removed, and whether performance during learning persists in the final fixed controller.
- What they did: A planar two-link arm reaching task in eight directions used velocity-dependent sideways forces, with online adaptation only in an external linear torque readout while feature parameters stayed fixed across CfC, echo-state, and RBF variants.
- The result: All six variants met the after-effect criterion and improved tracking; median normalized squared residuals ranged from 0.0097 to 0.383, and with updates off all final readouts retained useful compensation, while LMS variants had median fixed-readout error ratios of 2.28–5.87 and RLS stayed nearly unchanged.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.