A Wearable System Prototype for Activity-Contextualized Knee-Joint Health Assessment via Acoustic Emissions

Töreyin, Hakan ;   Teague, Caitlin ;   Hersek, Sinan ;   Millard-Stafford, Mindy ;   Jones, Michael ;   Kogler, Geza ;   Sawka, Michael ;   Inan, Omer

A wearable system prototype for activity-contextualized assessment of knee joint health is presented. The system can accurately detect airborne acoustic emissions as well as skin vibrations associated with movement at the knee joint. Microphone signals are amplified and processed using custom analog circuits, then sampled and encoded at 100 kS/s and 12-bits per channel, respectively. Absolute knee angle computation and basic activity classification is performed in hardware by capturing data from single axis gyroscope and accelerometer pairs (12-bits and 1 kS/s per sensor) placed on the shank and the thigh. Data logging and processing are conducted with a commercially available embedded processing tool, which has a form factor of 13.7 x 8.6 x 2.5 cm. Several hours of continuous recording time is possible using standard Lithium Ion Polymer batteries conducive to a wearable form factor. Enabling knee-joint sound detection in ambulatory subjects for the first time, the proposed wearable system has the potential of advancing the scientific understanding of how longitudinal changes in knee joint sounds correlate to changes in joint health.