Mission Planning for Small UAV Systems in Uncertain Environments

Abstract

The DLR Institute of Flight Systems is developing the experimental flying platform ARTIS (Autonomous Rotorcraft Testbed for Intelligent Systems), a fully autonomous miniature helicopter. The main concern of the project is the development of new subsystems and algorithms for autonomous intelligent functions. This paper presents an overview of the concepts related to mission planning and management for this research platform. The underlying problem of mission planning for a helicopter differs from high-altitude systems by the operation close to the ground, among known or unknown obstacles. For the global path planning in a three-dimensional space, previously known information about the environment is used. The path planning is based on the visibility graph concept, which returns length-optimal paths in a plane. To account for the whole 3D space, this algorithm is run in several fixed-height layers, and then connected to the start and finish points, generating near-optimal paths in a very short time. Subsequently, the path planner is completed by a real-time collision avoidance planner. Upon detection of unknown, static obstacles, the global path can be adapted in short time. The arithmetic techniques of this approach are further explained through examples. These mission planning components are integrated in the ARTIS ground control station software and thus can be used to plan and monitor mission in flight tests. Furthermore, the integration into the flying system and the correspondent problems are commented, as well as knowledge from flight experiences.