The thesis aimed to study, experiment and evaluate a prototype of a conversational engine for driver-to-vehicle interaction in the context of a connected car.

Specifically, the work aimed at developing a multimodal conversational application, based on cloud services, allowing to sustain conversations about the state of the vehicle and a subset of it driving and accessory functions, according to a set of predefined use cases.

The main steps of the thesis included:

1. analysis and selection of a suitable conversational cloud platform;
2. starting from the definition of the use cases, identify: the syntactic and linguistic assets used in the conversation, the state of the vehicle that needs to be modeled, the vehicle "physical" UI elements that may take part in the conversation.
3. design, implementation and test of the the conversational application and of the minimal vehicle model
4. integration of a voice front-end and experimentation with a set of users.