DocumentCode :
3657435
Title :
Wind feed-forward control of a USV
Author :
Huajin Qu;Edoardo I. Sarda;Ivan R. Bertaska;Karl D. von Ellenrieder
Author_Institution :
Department of Ocean &
fYear :
2015
fDate :
5/1/2015 12:00:00 AM
Firstpage :
1
Lastpage :
10
Abstract :
Since Unmanned Surface Vehicles (USVs) are relatively light weight and have a large windage area, they are sensitive to wind disturbances, especially during station-keeping operations. While robust feedback controllers can be used to attenuate slowly changing wind disturbances, they have difficulty compensating for the rapid variation of wind speed and direction. Control and sensing strategies were developed to mitigate wind effects on a 4.9 meter long, 180 kg USV. A wind feed-forward controller was designed using an anemometer to measure the apparent wind speed/direction and a wind model to estimate the wind-induced forces and moments on the USV in real-time. Freely drifting tests were performed to evaluate the impact of the wind on an unpowered USV. The vehicle was blown more than 20 meters off station in 5 minutes in mean wind speeds of about 2m/s. Experiments also showed that the response time of the USV to wind gusts is about 5 seconds. The wind feedforward feature was added to different station-keeping feedback controllers, namely a PD controller, a robust backstepping controller, and a sliding mode controller. Station keeping tests of those controllers without/ with wind feedforward controller were conducted and repeated for 5 times. The experimental results showed that the small size of USV can gain significant benefits from the inclusion of wind feedforward feature in both position and heading by an average of 75%, and 46% respectively in 5 sets of station keeping experiments.
Keywords :
"Wind speed","Feedforward neural networks","Vehicles","Fluid flow measurement","Mathematical model","Adaptive control","Control systems"
Publisher :
ieee
Conference_Titel :
OCEANS 2015 - Genova
Type :
conf
DOI :
10.1109/OCEANS-Genova.2015.7271438
Filename :
7271438
Link To Document :
بازگشت