Turn at the roundabout:A practical assessment of spatialrepresentations in two different GPS interfacesfrom a pedestrian’s perspective

Abstract: The role played by landmarks and 3D visualisations in navigation and wayfinding wasinvestigated from a pedestrian’s perspective. Two GPS units, a Garmin GPSMap 76CSand an hp iPAQ running Navigon MobileNavigator (Navigator) were assessed todetermine the usability of their interfaces. During the evaluation, particular attention waspaid to the use of landmarks and 3D representations in the maps displayed and theroutes directed to the user. Both systems used some landmarks in their display of theroute. The role that these landmarks played in describing and assisting in navigation toa pedestrian, and thus the value and benefits of their inclusion, was found to vary.The Garmin interface, a dedicated GPS unit, displayed landmarks when the maps wereviewed at a scale greater than 1:20 000. The landmarks displayed included restaurants,fast food outlets and service stations. Whilst represented on maps, these landmarkswere not referred to in the given wayfinding direction. The other tested interface,Navigator, was a PDA based interface, with the hardware capable of other tasksunrelated to GPS navigation. The Navigator interface also displayed landmarks,including fast food outlets, restaurants and car dealerships. The only landmarks to beexplicitly referred to in either interface were the roundabouts in the Navigator interface,where any instructions requiring a turn at a roundabout would explicitly mention that theturn was to be made at the roundabout. Both systems would only refer to buildingsdisplayed if one of those buildings displayed was the user’s destination. Thus, the otherbuilding landmarks represented in the maps served only as a form of reassurance forthe user, and using these to assist in navigation relied upon interpretation from the user.The landmarks displayed were found to be most useful when easily identified on thestreet and were in prominent positions. A clear representation on screen, which did notrely upon labelling the landmarks, was found to be of significant benefit. This both madean instantly recognisable representation and kept the interface’s display free fromclutter. Systems could be improved by future development to display landmarks at morecritical points along a particular route, and explicitly referring to these in the givendirections.Another distinct limitation in effectiveness was identified in the use of 3-dimensionalrepresentations, with neither system reflecting the 3-dimensional space well. Bothinterfaces represented their 3-dimensional views without any building appearancedetails. Excluding landmarks is likely to be a cost based decision made by the softwaredevelopers. There would be significant expense involved in creating and updating eachbuilding along a particular route. However, excluding building detail made the spacesurrounding the user appear open and empty in the 3D viewing modes. This poorlyrepresented the area in an urban setting, where buildings play a large role in definingspace. Improving this representation, to reflect a more closed in and confined space,was also identified as an opportunity for future development for both tested interfaces.