Using Real-Time Smoothness Metrics to Deliver Haptic Performance Cues for a Dexterous Task

[摘要] Surgeons representing a number of domains, including endovascular surgery, haveexpressed a desire for objective performance measures and feedback mechanisms asa means to improve training and assessment of skill. Metrics that quantify smoothnessof tool movements have recently been shown to exhibit strong correlations withendovascular surgical skill. This thesis explores the applicability and implementationof two of these smoothness metrics as a method for providing real-time performancefeedback via a haptic interface. The algorithms for spectral arc length and submovementcount are modi ed for near real-time computation. They are then mapped tovibrotactile cues to serve as a feedback when subjects complete a dexterous, mirrortracing task. This task is developed and veri ed as a suitable proxy task for endovascularsurgery and serves as a platform to test performance assessment and feedbackbased on movement smoothness analysis. Submovement count is shown to have limitedviability as it is constrained by its computationally expensive nature. However,spectral arc length based feedback is shown to have a substantial positive e ect onperformance in terms of both accuracy and completion time.

[发布机构] Rice University

[效力级别] Smoothness