Development of the HRL Route Navigation Dialogue System

Development of the HRL Route Navigation Dialogue System Robert Belvin HRL Laboatories, LLC 3011 Malibu Canyon Road Malibu, CA 90265 310-317-5799 rsBelvin@hrl.com Ron Burns HRL Laboatories, LLC 3011 Malibu Canyon Road Malibu, CA 90265 310-317-5445 rBurns@hrl.com Cheryl Hein HRL Laboatories, LLC 3011 Malibu Canyon Road Malibu, CA 90265 310-317-5489 cheryl@hrl.com ABSTRACT In this paper we report on our work on a prototype route navigation dialogue system for use in a vehicle. The system delivers spoken turn-by-turn directions, and has been developed to accept naturally phrased navigation queries, as part of our overall effort to create an in-vehicle information system which delivers information as requested while placing minimal cognitive load on the driver. Keywords Dialogue Systems, Discourse, Navigation, NLP, Pragmatics, Dialogue Manager 1. INTRODUCTION In this paper we report on our work on a spoken language navigation system which runs in real-time on a high-end laptop or PC, for use in a vehicle. We focus on issues in developing a system which can understand natural conversational queries and respond in such a way as to maximize ease of use for the driver. Because today’s technology has the potential to deliver massive amounts of information to automobiles, it is crucial to deliver this information in such way that the driver’s attention is not diverted from the primary task of safe driving. Our assumption has been that a dialogue system with a near-human conversational ability would place less of a cognitive load on the driver than one which behaves very differently than a human. We have implemented a testbed on which to develop and evaluate driver interfaces to navigation systems. Our approach is multi- modal and the interface will include a head-up display, steering hub controls, and spoken language, though it is only the latter modality that we report on here. We first discuss our development phases, and after this we provide an overview of our implementation, emphasizing the natural language processing aspects and application interface to the map databases. Next we provide results of our initial evaluation of the system, and finally we draw conclusions and summarize plans for future work. 2. DEVELOPMENT PHASES One can identify four distinct subproblems which must be solved for a navigation system: 1) the natural language navigation interface, 2) street name recognition, 3) the natural language destination entry interface given street name recognition, and 4) the map database interface. We have partitioned the problem and have phased our development to progressively implement solutions with increasing complexity. Navigation system implementation is complicated by the potential of having a very large street name vocabulary with many unusual and uncommon pronunciations with significant variations across speakers. The appropriate name space is dynamic since it depends on the location of the vehicle. Our initial system does not accept queries with proper street names. In addition, we assume separate destination entry and route planning systems, and that one or more routes have been loaded into the navigation system. The system relies on open dialogue to resolve the directions at any stage …

Content truncated to 500 words