SPINDLES+: An adaptive and personalized system for leg shake detection

For many people, the habit of leg shaking is an annoyance to themselves and people around them. In this work, we present SPINDLES+, a smartphone-based leg shaking detection application that detects and alerts users of leg shaking. In addition to being a low-power, low-latency, non-intrusive and high accuracy mobile application, SPINDLES+ extends the work of its predecessor, SPINDLES, by addressing the need for personalization in detecting a body motion that is unique to the individual. We show that SPINDLES+ is able to learn the characteristics that define an individual׳s leg shaking signature and allows the user to adjust the parameters of the detector, based on personal preference, to find an optimal point between accuracy, latency, and power consumption. Finally, we compare SPINDLES+ with related works to further illustrate improvements over the robustness and personalization aspects that SPINDLES+ affords. By providing leg shaking alerts in a timely manner, we hope to help users become more aware of and reduce the occurrence of this behavior.