Land Vehicle Control Using Continuous Gesture Recognition on Smartwatches

This work introduces a method for interaction between wearable devices and ground vehicles, utilizing continuous gesture recognition on a smartwatch to control an Arduino-based vehicle. Two interaction techniques were developed and compared: in the first, the vehicle moves while the user touches the screen; in the second, a double-tap initiates or stops the vehicle. Experiments included the application of SUS and NASA TLX questionnaires to assess usability and aspects such as mental load and satisfaction. Results indicate a significant user preference for the second technique, supported by SUS scores. Additionally, NASA TLX analysis reveals positive perceptions in dimensions of mental, physical, temporal load, performance, effort, and frustration. The distribution of preferences highlights that seven users favored technique two, two opted for technique one, and one did not express a preference. This study contributes to the development of intuitive and effective interactions between wearable devices and ground vehicles, exploring the potential of continuous gesture recognition.