Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Reactive Path Planning for an Autonomous model Sailboat

Author(s):

J. Wirz
Conference/Journal:

Master Thesis, FS15 (10397)
Abstract:

Guiding an autonomously moving vehicle from its current position to some target is referred to as path planning or navigation. Path planning adds a cognition level and deals with decision making and optimization of the path with respect to several possible goals. This thesis develops a path planning algorithm for an autonomous model sailboat. For this purpose, existing path planning and obstacle avoidance algorithms for autonomously moving vehicles are evaluated. Specifically, a potential field and cost based approach is implemented and tested in numerical simulations. The results from these simulations are used to develop a modified cost method, which improves and extends the existing methods. The newly developed method optimizes the obstacle avoidance, adds obstacle avoidance for moving obstacles and includes race tactical considerations. Furthermore, optimization of the run time of the algorithm allows it to be executed on embedded platforms with limited computational resources. Path planning for a sailboat is not trivial, since the physics of sailing induce an angular "no-sail" zone, where it is indeed not possible to sail. Further, the speed of a sailboat is directly related to the wind, which makes it nearly impossible to stop in front of obstacles and therefore requires careful and forward-looking navigation in order to avoid obstacles. As a second aspect, this thesis focuses on the development of a specialized hardware to detect and identify obstacles online. A one dimensional laser distance sensor is modified to create a distance map of the environment. A Kalman Filter based tracking algorithm tracks the obstacle positions and estimates the position and velocity of the potential obstacles. All algorithms developed in this thesis are implemented and validated on the test platform AEOLUS, a one meter long remote controlled sailboat.

Supervisors: Henrik Hesse, Sergio Grammatico, Roy Smith

Year:

2015
Type of Publication:

(12)Diploma/Master Thesis
Supervisor:

H. Hesse

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2015:IFA_5207
}
Permanent link