Abstract

Acoustic emission (AE) has been explored as a non invasive approach for knee assessment. However, existing methods based on a single sensor to capture knee AE lack the
spatial interpretability to determine the compartmental origin of detected AE events. This technical gap is addressed in this paper with a novel dual-sensor approach to achieve compartment specific assessment of the knee during dynamic movement. In addition to simultaneous acquisition of AE signals from medial and lateral knee regions, a static knee-tapping protocol was developed to generate impulsive AE signals from predefined anatomical locations, thereby providing essential ground truth for empirical characterization of inter-sensor features. Using arrival time delay, spectral correlation, energy decay and peak-amplitude ratio, a signal processing scheme was developed for identification of independent and co-detected AE hits as well as compartment classification. Experiments with healthy participants performing sit-stand-sit movements revealed distinct and reproducible compartment-level AE patterns, with the medial tibiofemoral compartment exhibiting the highest AE activity, followed by the lateral tibiofemoral and patellofemoral compartments, consistent with expected joint mechanics. Sensitivity analysis further confirmed the robustness of the classification criteria. Overall, the work establishes dynamic medial-lateral AE as a practical approach for compartment-specific knee assessment, with promising potential for clinical translation.