Towards Interaction, Disturbance and Fault Aware Flying Robot Swarms

Abstract

Visual navigation, mapping, obstacle avoidance and autonomous operation are becoming increasingly commercially available in flying robots. Recent research in control of flying robots has therefore shifted beyond sensor based trajectory tracking towards physical interaction and manipulation control. However, current research mainly focuses on specialized interaction cases under indoor or hovering conditions. Furthermore, the problem of physical interaction control for entire robot swarms operating under different control objectives is essentially unexplored terrain, as is the systematic treatment of disturbances and faults for both single and swarm systems, respectively. In this position paper, we argue that robust operation of interacting flying robots requires systematic handling of interactions and external inputs such as faults from individual robot to swarm level. For this, we introduce a scalable awareness methodology for interaction, disturbance and fault handling resulting in the awareness pipeline scheme. Another algorithmic key element for unification is the extension of well established methods from operational space and multipriority robot control to this system class, potentially leading to novel controls and skills of flying robot swarms.