SENSOR-BASED ONLINE PATH PLANNING FOR ROTORCRAFT LANDING

Abstract

This paper describes a reactive path planning strategy to avoid obstacles in partially known environments during the approach-to-landing phase of a manned helicopter. As experimental platform, the research rotorcraft Active Control Technology/Flying Helicopter Simulator (ACT/FHS) is used. This highly modified EC135 is equipped with a commercial forward-looking Light Detection and Ranging (LiDAR) sensor with a range of 1 km. During an approach to an unmapped landing site, geo-referenced LiDAR samples are acquired and combined with database information. The resulting representation of the environment is used for the generation of flight paths which are collision free, technically feasible and acceptable for pilots. Once new samples are collected by the LiDAR sensor, the environment map is updated in real time and the flight path is changed based on typical approach procedures if necessary. Due to experimental aspects, a manual trajectory following was used by providing the pilot with a "Tunnel-In-The-Sky" head down display including visual cues for spatial and speed guidance during the approach. Results show the applicability of the presented planning strategy based on five approaches made to Braunschweig airport in a flight trail conducted in 2015. Furthermore, the need for proper Human-Machine-Interface design is indicated in order to communicate decisions made by the system to the pilot, which exceeds the scope of this work.