Robust control of a vessel using camera feedback and extended Kalman filter

This paper presents the implementation of a robust position tracking control law for an underactuated autonomous surface vessel. The robust control law is developed based on the sliding mode approach where a first order sliding surface is defined in terms of surge and a second order one in terms of lateral motion tracking errors. The control law uses a simple three-degree-of-freedom planar vessel model with two actuator inputs. The vessel is a small model boat with two propellers in a small indoor pool. The position and orientation of the boat is measured using a camera and with two infrared diodes attached the front and end of the boat. An Extended Kalman Filter is designed to estimate all the unmeasured states and filter out the measurement noise. A computer with controller board processes the camera image, calculates the control inputs, and sends the control signal to a two channel wireless receiver on the vessel using a wireless transmitter. The relationship between the distorted camera image and the vessel actual position as well as the relation between motor input voltage and propeller force are derived through experimental calibrations. An experiment is performed where the vessel follows a predefined straight-line trajectory.