FutureNav The Design and Evaluation of a Strategic Navigation Interface for Semi-Autonomous Driving in Urban Environments

Abstract

Turn-by-turn GPS systems are designed for todays manual drive cars, focusing pri- marily on providing short-term instructions on steering the vehicle. In the future, semi-autonomous vehicles will take over much of the steering duty, and in the pro- cess, change the type of navigation information which is valuable to the driver. We identify 2 scenarios where information provided by the navigation interface could be valuable. The first is a challenge: to enhance the driver awareness of their surround- ings in the case they have to take over operation. The seconds is an opportunity: to provide strategic long-term control of the vehicle’s heading. In accordance with these 2 scenarios, we conducted a 50 person survey to study the information needs of drivers of semi-autonomous vehicles. The results indicate that these information needs center primarily around Time, Location, Route Conditions, and Self-Driving Parts. We also find that these vary between different trip purposes: social trips, home trips, and shopping trips. To test the validity of these findings, we design and construct a working prototype of a navigation interface, FutureNav, which provides this information and gives drivers high-level control over it. Testing the hypothesis that such an interface may increase driver awareness, we conduct an experiment in the real-world comparing FutureNav to Google Maps as a baseline, using standard measures for testing situation awareness on all 3 levels. We identify certain features of FutureNav, such as the ghost car,map landmarks, and the 3D viewpoint as po- tential sources of increased situation awareness. Finally, we reflect on our design decisions, and the learning from conducting in-the-wild experiments that focus on situation awareness, noting that such information may be useful for the design and testing of future navigation interfaces designed for semi-autonomous vehicles.