Information Technology Reference
In-Depth Information
Chapter 19
Unmanned Helicopter Control via Visual
Servoing with Occlusion Handling
Yasushi Iwatani, Kei Watanabe, and Koichi Hashimoto
Abstract.
This chapter presents an automatic take-off and landing system for an
unmanned helicopter using visual servo control. Cameras are placed on the ground,
and a static-eye configuration is set up. A vision-based control method with occlu-
sion handling is implemented. The reference for the altitude in take-offs and land-
ings is specialized to operate in ground effect. An experimental result demonstrates
that the helicopter is controlled accurately.
19.1
Introduction
Computer vision plays an important role for automatic control of unmanned aerial
vehicles and micro aerial vehicles [3, 4, 9]. In particular, cameras are useful to mea-
sure the position of a vehicle relative to the world reference frame. Vision-based
control is also powerful in take-offs and landings, since vision provides the precise
position relative to a given landing pad.
This chapter mainly focuses on automatic take-off and landing control of an un-
manned helicopter using vision sensors. Restricted to take-offs and landings, there
are two possibilities for camera configurations. In a hand-eye configuration, cameras
are mounted on the helicopter. In a static-eye configuration, cameras are placed on
the ground. Vision-based control in static-eye configurations is more robust to cali-
bration errors than hand-eye configurations [1]. Thus a static-eye configuration is set
up in this paper, while previous vision-based landing control systems for unmanned
helicopters have hand-eye configurations [5, 6, 7].
In this chapter, visual servo control with occlusion handling proposed in [2] is
implemented to avoid the occlusion problem caused by the rotor motion. The im-
plemented method first chooses a set of correctly extracted image features in the
selection step. It then estimates all the image features from the correctly extracted
Search WWH ::
Custom Search