Robust natural language processing for urban trip planning.

Navigating large, urban transportation networks is a complicated task. A user needs to negotiate the available modes of transportation, their schedules, and how they are interconnected. In this paper we present a Natural Language interface for trip planning in complex multimodal urban transportation networks. Our objective is to provide robust understanding of complex requests while giving the user flexibility in their language. We designed TRANQUYL, a Transportation Query Language for trip planning; we developed a user-centric ontology, that defines the concepts covered by the interface and allows for a broad vocabulary. NL2TRANQUYL, the software system built on these foundations, translates English requests into formal TRANQUYL queries. Through a detailed intrinsic evaluation, we show that NL2TRANQUYL is highly accurate and robust with respect to paraphrasing of requests as well as handling fragmented or telegraphic requests.