Prior research has shown piezoelectric vibration sensors can detect sail luffing, the acceleration of a flexible sail out of its normal wing state caused by a momentary reversal of the air pressure gradient over the sail. Luffing decreases boat performance by reducing the lift generated by the sail. Yet detecting sail luffing in flexible sails for robotic sailboats is still challenging. This paper presents three methods of sensing characteristics of a luff – air pressure differential which causes the luff, the acceleration of the sail as the luff occurs, and the influence of motion and acceleration of a luffing sail on members placed on the sail. We assess three different sensor types based on cost, ease of use, complexity of electrical interface, power consumption, accuracy of the sensor and amount of noise in sensor readings. To classify the most effective sensor for a given set of constraints, a multifaceted analysis has been performed with a piezoelectric vibration sensor, an acceleration sensor, and a gas pressure sensor. The accuracy and precision of each sensor at sensing sail luffing is evaluated by comparing the sensor output with a plot of the position of a single point on the sail through time generated with computer vision.

Current robotic sailing relies on sensing wind direction and moving the sails to a position that is appropriate for that relative wind angle. The research to date shows that a correctly tuned sail in the classic wing shape is essential for maximum speed over water. This paper relates research on sensors used to determine when sail trim is incorrect. With improper sail trim, the sail luffs. This luffing produces turbulence which reduces the efficiency of the sail. By instrumenting the sail with sensors to detect when sails begin to luff, the robot can determine when the sail is improperly trimmed and, potentially, take corrective action.

The robotic revolution dramatically increased the number of tasks which robots perform. It continues to do so, especially within the maritime industry, where robotic applications are being developed across a wide variety of platforms from power to sail. One challenge faced when constructing sailing robots is the inability to determine the state of the sail if there is no human at the helm watching and reacting to constantly changing wind conditions. Traditional, flexible sails begin to “luff” or flap in the wind if are not set correctly. With this comes a decrease in sail efficiency and hull speed because the sails no longer generate maximum lift. In the worse case, the boat reaches a state of being in “irons” where it stalls in the water, bow to the wind. In previous research I showed sensors could be placed on the sail to detect this luffing. In this research, I present a quantitative comparison of three different methods of sensing sail luffing: piezoelectric vibration, air pressure, and acceleration. The relative effectiveness of these methods is compared against position data obtained using computer vision during sets of binary tests where the sail can be considered to luff or to hold its shape. Results are promising and show that each sensor is able to detect sail luffing. However, there are also clear use cases where one sensor is preferable.