Low-Cost Tactile Localization on Multigeometry Surfaces via Hybrid Signal Analysis

Enabling affordable, high-precision tactile sensing on robots remains a key challenge for safe and intuitive human-robot interaction. Conventional approaches often rely on expensive electronic skins or require complex signal processing with high sampling rates and precise time synchronization—barriers to scalable deployment. Precise tactile localization on robotic shells using low-cost sensors is challenging. This paper proposes a novel method that integrates Signal Intensity Ratio (SIR) and waveform correlation for accurate impact localization on both planar and curved surface. A key innovation is that the SIR, measured between sensor-pairs, is invariant to the intensity of the impact, eliminating the common need for high sampling rates and time synchronization. The waveform correlation serves as supplementary criterion, together overcoming accuracy degradation from wave dispersion and reflections. Experimental validation on ethylene-vinyl acetate foam panels, using four low-cost microphones (44.1kHz), showed mean errors of 1.55 mm (flat) and 1.84 mm (curved), with 90th percentile errors below 3.16 mm and 3.61 mm, respectively. Furthermore, preliminary validation was extended to different materials, impact types, and surface geometries. This work provides a cost-effective pathway for adding tactile sensing to robotic shells, with potential applications in collision detection and interactive interfaces.