https://sam.ensam.eu:443 The DSpace digital repository system captures, stores, indexes, preserves, and distributes digital research material. Wed, 13 May 2026 16:42:21 GMT 2026-05-13T16:42:21Z http://hdl.handle.net/10985/15096 Methodology for optimal wind vane design KERHASCOET, Hugo; LAURENT, Johann; CERQUEUS, Audrey; SEVAUX, Marc; SENN, Eric; HAUVILLE, Frederic; CONEAU, Raphael Measurements of wind direction are sought after by a multitude of professionals in many different domains. Whether to recalibrate meteorological models or simply for leisure activities, the demand for accurate and responsive wind measurements is widespread. This study was motivated by the need to improve the responsiveness of direction measurements on yachts. Here we argue that the ideal form factor of the wind sensor can be determined using digital tools, rather than empirically, with the aim of improving the mechanical response of the wind vane. Then we present the results obtained by applying a predictive filter method tailored to the specified form factor. We have developed and experimentally validate a mathematical model describing the dynamic behavior of a wind vane. This model is then used to determine the form factor of the vane that will give the best possible response to perturbations it will encounter. To do so we use operational research tools, specifying the mechanical characteristics of the vane and by providing the future use conditions of the sensor, in the form of a wind speed spectral density. The design built from this optimization methodology helps reduce the response time of the vane by 44% compared to to designs currently in use. We then work on digital signal processing by using a predictive filter which takes into account the dynamic characteristics of the vane previously determined by the mathematical model. This step vastly improves the quality and sensitivity of the signal, leading to another reduction in response time of 83%. This brings the total decrease in response time at 90%. There is therefore not only an improvement in the quality of wind direction measurements, but also with respect to the set of data that is derived from this information. In the context of single-handed racing boats, the performance of the automatic pilot directly benefits from this improvement in responsiveness. Fri, 01 Jan 2016 00:00:00 GMT http://hdl.handle.net/10985/15096 2016-01-01T00:00:00Z KERHASCOET, Hugo LAURENT, Johann CERQUEUS, Audrey SEVAUX, Marc SENN, Eric HAUVILLE, Frederic CONEAU, Raphael Measurements of wind direction are sought after by a multitude of professionals in many different domains. Whether to recalibrate meteorological models or simply for leisure activities, the demand for accurate and responsive wind measurements is widespread. This study was motivated by the need to improve the responsiveness of direction measurements on yachts. Here we argue that the ideal form factor of the wind sensor can be determined using digital tools, rather than empirically, with the aim of improving the mechanical response of the wind vane. Then we present the results obtained by applying a predictive filter method tailored to the specified form factor. We have developed and experimentally validate a mathematical model describing the dynamic behavior of a wind vane. This model is then used to determine the form factor of the vane that will give the best possible response to perturbations it will encounter. To do so we use operational research tools, specifying the mechanical characteristics of the vane and by providing the future use conditions of the sensor, in the form of a wind speed spectral density. The design built from this optimization methodology helps reduce the response time of the vane by 44% compared to to designs currently in use. We then work on digital signal processing by using a predictive filter which takes into account the dynamic characteristics of the vane previously determined by the mathematical model. This step vastly improves the quality and sensitivity of the signal, leading to another reduction in response time of 83%. This brings the total decrease in response time at 90%. There is therefore not only an improvement in the quality of wind direction measurements, but also with respect to the set of data that is derived from this information. In the context of single-handed racing boats, the performance of the automatic pilot directly benefits from this improvement in responsiveness. http://hdl.handle.net/10985/15095 Speedometer Fault Detection and GNSS Fusion using Kalman Filters KERHASCOET, Hugo; LAURENT, Johann; SENN, Eric; HAUVILLE, Frederic Navigation systems used in racing boats require sensors to be more and more sophisticated in order to obtain accurate information in real time. To meet the need for accuracy of the surface speed measurement, the mechanical sensor paddle wheel has been replaced by the ultrasonic sensor. This ultrasonic sensor measures the water speed precisely and with very good linearity. Furthermore, by its principle of operation, it measures the water flow several centimetres from the sensor, which puts it outside the boundary layer, the region close to the hull where the flow is disturbed. However, this sensor has several drawbacks: it is quite sensitive and if the flow contains too many air bubbles, the sensor picks them up, which can happen quite frequently on boat with a planing hull. Another limitation of this sensor is its low frequency measurement rate. In this paper we explain the techniques used based on Kalman filters to address these shortcomings, firstly by identifying the inaccurate measurements caused by inadvertent dropouts, then by improving the useful sensor frequency with GNSS data fusion. Fri, 01 Jan 2016 00:00:00 GMT http://hdl.handle.net/10985/15095 2016-01-01T00:00:00Z KERHASCOET, Hugo LAURENT, Johann SENN, Eric HAUVILLE, Frederic Navigation systems used in racing boats require sensors to be more and more sophisticated in order to obtain accurate information in real time. To meet the need for accuracy of the surface speed measurement, the mechanical sensor paddle wheel has been replaced by the ultrasonic sensor. This ultrasonic sensor measures the water speed precisely and with very good linearity. Furthermore, by its principle of operation, it measures the water flow several centimetres from the sensor, which puts it outside the boundary layer, the region close to the hull where the flow is disturbed. However, this sensor has several drawbacks: it is quite sensitive and if the flow contains too many air bubbles, the sensor picks them up, which can happen quite frequently on boat with a planing hull. Another limitation of this sensor is its low frequency measurement rate. In this paper we explain the techniques used based on Kalman filters to address these shortcomings, firstly by identifying the inaccurate measurements caused by inadvertent dropouts, then by improving the useful sensor frequency with GNSS data fusion. http://hdl.handle.net/10985/15041 Sensor Fault Detection and Signal Improvement using Predictive Filters KERHASCOET, Hugo; MERIEN, P; LAURENT, Johann; SENN, Eric; HAUVILLE, Frederic Navigation systems used in racing boats require sensors to be more and more sophisticated in order to obtain accurate information in real time. To meet the need for accuracy of the surface speed measurement, the mechanical sensor paddle wheel has been replaced by the ultrasonic sensor. This ultrasonic sensor measures the water speed precisely and with very good linearity. Furthermore, by its principle of operation, it measures the water flow several centimetres from the sensor, which puts it outside the boundary layer, the region close to the hull where the flow is disturbed. However, this sensor has several drawbacks: it is quite sensitive and if the flow contains too many air bubbles, the sensor picks them up, which can happen quite frequently on boat with a planing hull. Another limitation of this sensor is its low frequency measurement rate. In this paper, we explain the techniques used based on Kalman filters to address these shortcomings, firstly by identifying the inaccurate measurements caused by inadvertent dropouts, then by improving the useful sensor frequency with GNSS data fusion. Sun, 01 Jan 2017 00:00:00 GMT http://hdl.handle.net/10985/15041 2017-01-01T00:00:00Z KERHASCOET, Hugo MERIEN, P LAURENT, Johann SENN, Eric HAUVILLE, Frederic Navigation systems used in racing boats require sensors to be more and more sophisticated in order to obtain accurate information in real time. To meet the need for accuracy of the surface speed measurement, the mechanical sensor paddle wheel has been replaced by the ultrasonic sensor. This ultrasonic sensor measures the water speed precisely and with very good linearity. Furthermore, by its principle of operation, it measures the water flow several centimetres from the sensor, which puts it outside the boundary layer, the region close to the hull where the flow is disturbed. However, this sensor has several drawbacks: it is quite sensitive and if the flow contains too many air bubbles, the sensor picks them up, which can happen quite frequently on boat with a planing hull. Another limitation of this sensor is its low frequency measurement rate. In this paper, we explain the techniques used based on Kalman filters to address these shortcomings, firstly by identifying the inaccurate measurements caused by inadvertent dropouts, then by improving the useful sensor frequency with GNSS data fusion.