Structured Exploratory Strategies in Volumetric Haptic Palpation of Virtual Phantom Tissues with Nodules Revealed by Trajectory-Based AI Inference

Digitalizing manual tissue examination requires also capturing the tactile expertise of human palpation. We present a haptic framework for studying and modeling exploratory strategies during volumetric tissue interaction, using an ex vivo-inspired virtual environment with uncertainty driven by coupled nodule size and depth. From pilot haptic trajectory data collected during nodule search tasks, LSTM models identify Global Scanning and Local Refinement behaviors and their transitions. Localized probing precedes improved detection-weighted localization, highlighting a trade-off between exploration and precision. These findings reveal structured exploratory strategies in haptic palpation and support encoding tactile expertise for haptic training systems and autonomous robotic palpation.