Publication: Diseño de un estimador basado en redes neuronales artificiales para caracterizar la frenada de un vehículo
Loading...
Identifiers
Publication date
2021-11
Defense date
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Fundación Técnica Industrial
Impact
Export
Abstract
Hoy en día, los vehículos inteligentes están equipados con interfaces de usuario muy avanzadas que pueden reaccionar ante las decisiones y necesidades del conductor. De esta manera mejora la dinámica vehicular según la situación de conducción evitando posibles colisiones fortuitas. Esta tecnología permite tomar decisiones sobre la conducción con un nivel de resolución superior al humano a la hora de, por ejemplo, realizar una frenada de emergencia ante un imprevisto. Con este trabajo se pretende que, en un futuro, se pueda gobernar el sistema de frenado de manera automática y que el vehículo pueda reproducir los hábitos y las formas de actuación de los conductores durante la maniobra de frenada de un vehículo, pero corrigiendo los posibles fallos humanos ligados a distracciones, falta de visibilidad o tiempos de reacción. Un sistema de frenado inteligente ha de ser capaz de obtener información de los diferentes sensores embarcados en el vehículo, procesar los datos obtenidos a través de los mismos y transformarlos en información útil para el control activo del automóvil. Con el vehículo instrumentado con los sensores capaces de proporcionar en tiempo real información sobre el sistema de frenado se llevan a cabo una serie de ensayos. Los datos obtenidos experimentalmente durante la realización de los ensayos se utilizan para el diseño de un estimador mediante la técnica de redes neuronales artificiales. Se estimarán las variables medidas por los sensores con el fin de caracterizar una frenada.
Today, smart vehicles are equipped with highly advanced user interfaces that can react to the driver’s decisions and needs. In this way, it improves vehicle dynamics according to the driving situation and avoiding possible accidental collisions. This technology makes it possible to make decisions about driving with a level of resolution higher than human, for example, carrying out an emergency braking in the event of an unforeseen event. With this work it is intended that, in the future, the braking system can be governed automatically and that the vehicle can reproduce drivers’ habits and forms of action during the braking maneuver of a vehicle, but correcting the possible human failures linked to distractions, lack of visibility and reaction times. An intelligent braking system must be able to obtain information from the different sensors on board the vehicle, process the data obtained through them and transform them into useful information for the active control of the car. With the vehicle instrumented with sensors capable of providing real-time information about the braking system, a series of tests are carried out. The data obtained experimentally during the tests are used to design an estimator using the artificial neural network technique. The variables measured by the sensors will be estimated in order to characterize a braking.
Today, smart vehicles are equipped with highly advanced user interfaces that can react to the driver’s decisions and needs. In this way, it improves vehicle dynamics according to the driving situation and avoiding possible accidental collisions. This technology makes it possible to make decisions about driving with a level of resolution higher than human, for example, carrying out an emergency braking in the event of an unforeseen event. With this work it is intended that, in the future, the braking system can be governed automatically and that the vehicle can reproduce drivers’ habits and forms of action during the braking maneuver of a vehicle, but correcting the possible human failures linked to distractions, lack of visibility and reaction times. An intelligent braking system must be able to obtain information from the different sensors on board the vehicle, process the data obtained through them and transform them into useful information for the active control of the car. With the vehicle instrumented with sensors capable of providing real-time information about the braking system, a series of tests are carried out. The data obtained experimentally during the tests are used to design an estimator using the artificial neural network technique. The variables measured by the sensors will be estimated in order to characterize a braking.
Description
Keywords
Sensor de presión, Redes neuronales artificiales, Estimación de la presión del sistema de frenado, Pressure sensor, Artificial neural network, Brake pressure estimation
Bibliographic citation
Garrosa, M., Olmeda, E., Sanz Sánchez, S. & Díaz, V. (2021). Diseño de un estimador basado en redes neuronales artificiales para caracterizar la frenada de un vehículo. Técnica Industrial, 330, 70-81.