Brain Computer Interface System for UAV Target Searching

Abstract

This paper proposes a Brain Computer Interface (BCI) system for unmanned aerial vehicle (UAV) indoor space target searching. The monocular vision navigation subsystem provides three-dimensional (3D) space feasible flight directions. To realize four-class motor imagery (MI) electroencephalogram (EEG) features extraction (left/right-hand, feet and tongue), the decision subsystem adopts a mathematical paradigm combining “one versus one” (OVO) and “one versus rest” (OVR) to traditional Common Spatial Pattern (CSP). Simultaneously, it utilizes a 2-layer hierarchical support vector machine (HSVM) to realize the feature classification and final feasible flight direction decision. The actual indoor space target searching experiment verifies that this BCI system has good adaptability.