Wavelet-Based Gesture Recognition Method for Human-Machine Interaction in Aviation

Human-machine interaction (HMI) has been an important research area in many scientific fields for many years with the development of sensor technologies. With the general use of unmanned aerial vehicles (UAVs) in recent years, the gesture recognition concept for HMI has become a considerable study in UAVs. They can be used in different domains and for different purposes. The main motivation for this paper is to create as simple a human-machine interface design for UAVs as possible. The vector-based features can be arranged in low memory size and the low-complexity decision process on them can be aimed. The gesture recognition problem for HMI in UAVs can be modeled as a pattern recognition problem in a multi-sensor system with vectorial data from various types of sensors. The speed of the action and the magnitude of the action can be different even within the same class of gesture action. In other words, user-independent gesture action recognition is difficult due to large intra-class differences in gesture action patterns. This paper introduces the wavelet-based vectorial feature extraction and the discriminant function-based decision in the supervised-based reduced vectorial dimension. According to experimental studies, the promising method is put forward for gesture recognition in the control of UAVs.