Publisher's Synopsis
Bachelor Thesis from the year 2015 in the subject Computer Science - Miscellaneous, grade: 1.0, Technical University of Darmstadt, language: English, abstract: Traffic safety is a major concern, especially for non-motorized traffic participants. According to various studies many accidents with cyclists, pedestrians, and athletes are caused due to insufficient visibility in traffic. To increase their visibility, we developed a wearable device which combines a lightweight sensor platform with a RGB LED display to give feedback to the surrounding traffic. This thesis presents a placing recognition for wearable devices, which allows autonomous placement detection utilizing an accelerometer. The data received by placement detection is then used to detect implicit and explicit gestures as a context aware system featuring a brake detection for back mounted devices and a turn indication for arm mounted devices. These gestures can be used to increase the cyclists safety, by using feedback modalities on RGB LED display used to communicate the intended actions to the following traffic. In order to archive low power consumption and a long battery run time, the presented algorithms for gesture recognition are optimized for low performance microcontrollers.
