Development of a speed profile prediction algorithm based on navigation data for energy management optimization

This master thesis work wants to develop and implement an on-board algorithm for the speed prediction of a connected vehicle along a given route, by elaborating real-time navigation data. In particular, information about the legal speed limits, the traffic density and the presence of the so-called "stop events", such as traffic lights and roundabouts, are sent to the vehicle by the map service provider and they constitute the input for the algorithm. The algorithm allows you to select one out of three different driver type (quiet, normal, aggressive): this choice, together with the performance parameters of the vehicle, influences the acceleration and braking phases of the prediction. Once the prediction is generated, it constitutes an input for the predictive ADAS functions and energy management functions. The work has been split in three main phases: development, calibration end validation. During the first phase, the logics of the algorithm have been implemented by means of a Simulink block in order to be included into the Hybrid Control Unit at the HiL. Subsequently, the calibration of the key parameters took place by means of real speed profile analysis. In the end, the behavior of the algorithm has been investigated studying its response in different scenarios.