Urban Positioning on a Smartphone: Real-time Shadow Matching Using GNSS and 3D City Models

TLDR: This paper describes the first real-time implementation of shadow matching on a smartphone capable of receiving both GPS and GLONASS and reports the first ever demonstration of any 3D-model-aided GNSS positioning technique in real time, as opposed to using recorded GNSS data.

Abstract: The performance of global navigation satellite system (GNSS) user equipment in urban canyons is particularly poor in the cross-street direction. This is because more signals are blocked by buildings in the cross-street direction than along the street [1]. To address this problem, shadow matching has been proposed to improve cross-street positioning from street-level to lane-level (meters-level) accuracy using 3D city models. This is a new positioning method that uses the city model to predict which satellites are visible from different locations and then compares this with the measured satellite visibility to determine position [2]. In previous work, we have demonstrated shadow matching using GPS and GLONASS data recorded using a geodetic GNSS receiver in Central London, achieving a cross-street position accuracy within 5m 89% of the time [3]. This paper describes the first real-time implementation of shadow matching on a smartphone capable of receiving both GPS and GLONASS. The typical processing time for the system to provide a solution was between 1 and 2 seconds. On average, the cross-street position accuracy from shadow matching was a factor of four better than the phone’s conventional GNSS position solution. A number of groups have also used 3D city models to predict and, in some cases, correct non-line-of-sight reception [4-6]. However, to our knowledge, this paper reports the first ever demonstration of any 3D-model-aided GNSS positioning technique in real time, as opposed to using recorded GNSS data. When it comes to real-time positioning on a smartphone, various obstacles exist including lower-grade GNSS receivers, limited availability of computational power, memory, and battery power. To tackle these problems, in this work, an efficient smartphone-based shadow-matching positioning system was designed. The system was then implemented in an app (i.e. application or software) on the Android operating system, the most common operating system for smartphones. The app has been developed in Java using Eclipse, a software development environment (SDE). It was built on Standard Android platform 4.0.3, using the Android Application programming interface (API) to retrieve information from the GNSS chip. The new positioning system does not require any additional hardware or real-time rendering of 3D scenes. Instead, a grid of building boundaries is computed in advance and stored within the phone. This grid could also be downloaded from the network on demand. Shadow matching is therefore both power-efficient and cost-effective. Experimental testing was performed in Central London using a Samsung Galaxy S3 smartphone. This receives both GPS and GLONASS satellites and has an assisted GNSS (AGNSS) capability. A 3D city model of the Aldgate area of central London, supplied by ZMapping Ltd, was used. Four experimental locations with different building topologies were selected on Fenchurch Street, a dense urban area. Using the Android app developed in this work, real-time shadow-matching positioning was performed over 6 minutes at each site with a new position solution computed every 5 seconds using both GPS and GLONASS observations were used for real-time positioning. The measurement data was also recorded at 1-second intervals for later analysis. Various criteria are applied to access the new system and compare it with the conventional GNSS positioning results. The experimental results show that the proposed system outperforms the conventional GNSS positioning solution, reducing the mean absolute deviation of the cross-street positioning error from 14.81 m to 3.33 m, with a 77.5 percentage reduction. The feasibility of deploying the new system on a larger scale is also discussed from three perspectives: the availability of 3D city models and satellite information, data storage and transfer requirements, and demand from applications. This meters-level across-street accuracy in urban areas benefits a variety of applications from Intelligent Transportation Systems (ITS) and land navigation systems for automated lane identification to step-by-step guidance for the visually impaired and for tourists, location-based advertisement (LBA) for targeting suitable consumers and many other location-based services (LBS). The system is also expandable to work with Galileo and Beidou (Compass) in the future, with potentially improved performance. In the future, the shadow-matching system can be implemented on a smartphone, a PND, or other consumer-grade navigation device, as part of an intelligent positioning system [7], along with height-aided conventional GNSS positioning, and potentially other technologies, such as Wi-Fi and inertial sensors to give the best overall positioning performance. / References [1] Wang, L., Groves, P. D. & Ziebart, M. Multi-constellation GNSS Performance Evaluation for Urban Canyons Using Large Virtual Reality City Models. Journal of Navigation, July 2012. [2] Groves, P. D. 2011. Shadow Matching: A New GNSS Positioning Technique for Urban Canyons The Journal of Navigation, 64, pp417-430. [3] Wang, L., Groves, P. D. & Ziebart, M. K. GNSS Shadow Matching: Improving Urban Positioning Accuracy Using a 3D City Model with Optimized Visibility Prediction Scoring. ION GNSS 2012. [4] Obst, M., Bauer, S. and Wanielik, G. Urban Multipath Detection and mitigation with Dynamic 3D Maps for Reliable Land Vehicle Localization. IEEE/ION PLANS 2012. [5] Peyraud, S., Betaille, D., Renault, S., Ortiz, M., Mougel, F., Meizel, D. and Peyret, F. (2013) About Non-Line-Of-Sight Satellite Detection and Exclusion in a 3D Map-Aided Localization Algorithm. Sensors, Vol. 13, 2013, 829?847. [6] Bourdeau, A., M. Sahmoudi, and J.-Y. Tourneret, “Tight Integration of GNSS and a 3D City Model for Robust Positioning in Urban Canyons,” Proc. ION GNSS 2012. [7] Groves, P. D., Jiang, Z., Wang, L. & Ziebart, M. Intelligent Urban Positioning using Multi-Constellation GNSS with 3D Mapping and NLOS Signal Detection. ION GNSS 2012.