Behavior-based High Level Control of a VTOL UAV

Abstract

Light weight unmanned aerial vehicles are expected to have a high level of system autonomy during missions in urban terrain although their onboard computational and payload resources are quite limited. This work addresses the problem of providing such platforms with capabilities to plan and conduct missions under constrained onboard perception of the environment and with reduced on-board assistance. Two behaviors are proposed - Fly Home, and Search and Track Object - which can operate in the absence of an online path planner and without an on-board world model. The approach focuses on reorganizing and adapting tasks of a-priori mission plans. As a result, even under these constrained planning conditions, each high level behavior implements mission critical features and reduces the workload of an UAV operator.