Position error reduction in a mechanical tracking linkage for arthroscopic hip surgery.

PURPOSE: Position tracking is an important aspect of many computer-aided surgical techniques. Given the obstacles of current optical, electromagnetic, and mechanical systems for medical applications, this work investigates error reduction in a new mechanical tracking system developed for arthroscopic hip surgery. This new tracking linkage addresses the current contradictory requirements of a thin, small linkage for ease of surgical use and a large, bulky linkage for increased accuracy by using (1) kinematic redundancy and (2) data averaging and curve fitting.
METHOD: To reduce the position error in the proposed mechanical tracking linkage, four numerical techniques were applied to data from the linkage. Two averaging techniques and two curve fitting techniques were investigated. After implementing each numerical technique, error testing was performed to quantify improvement in performance.
RESULTS: While the simple average and moving average techniques lowered the error by over 30%, these two methods were undesirable for the overall system performance. The lowest error was measured by the linear curve prediction method. This technique measured 0.552 mm of error, roughly half of the error measured by the control case. The quadratic prediction method reduced the error by 35% and had the lowest standard deviation in the measurements.
CONCLUSION: The linear prediction technique was able to significantly lower the error measured by a kinematically redundant mechanical tracking linkage. While further testing is still necessary, this data suggests that a thin, small mechanical tracking linkage could achieve a high level of accuracy to be an appropriate choice for a surgical tracking application.