Initial Study of Low Cost Crepitus Joint Degradation Finding Using Acoustic Localization

Current methods of joint diagnosis depend on invasive, radiative methods, large set ups and requiring skilled operators, which are still not able to assess the joint during motion. Replacing these methods are acoustic methods, which are only able to diagnose joint health based on the characteristics sound emitted by a moving joint. This paper looks into informing the medical practitioner of the locations of joint damage instead, hence the feasibility of localizing upon the damaged part of the joint, which is represented by a playback of a recording of crepitus sounds emitted from a human joint in 2 dimensional settings. Three microphones, an earphone and a data acquisition device were used to implement the testing. Using the time-difference-of-arrival (TDOA) between the 3 sensors, the Angle-of-Arrival (AOA) method was used to attain the initial localization coordinates. These coordinates were then fed into a gradient descent algorithm to reduce the positioning error of the AOA method, yielding the final position results. Results show that the developed system is able to achieve localization errors of less than 1.0 cm.