Medical simulators depend on realistic haptic feedback to train practitioners, but biological tissues are viscoelastic, exhibiting behaviors such as creep and stress relaxation. These properties are difficult to model accurately, yet essential for believable touch sensations.
Fractional-order models offer an effective way to represent such viscoelastic dynamics, but their parameters must be carefully estimated. The new work applies active learning to this parameter-estimation problem, aiming to make the process more efficient and accurate.
The approach is designed to improve the realism of haptic rendering in simulators, potentially leading to better training outcomes. The abstract does not provide details on the active learning algorithm or validation results, so the specific gains remain unspecified.