Richard Brunauer

Why GPS makes distances bigger than they are

ABSTRACT Global navigation satellite systems such as the Global Positioning System (GPS) is one of the most important sensors for movement analysis. GPS is widely used to record the trajectories of vehicles, animals and human beings. However, all GPS movement data are affected by both measurement and interpolation errors. In this article we show that measurement error causes a systematic bias in distances recorded with a GPS; the distance between two points recorded with a GPS is – on average – bigger than the true distance between these points. This systematic ‘overestimation of distance’ becomes relevant if the influence of interpolation error can be neglected, which in practice is the case for movement sampled at high frequencies. We provide a mathematical explanation of this phenomenon and illustrate that it functionally depends on the autocorrelation of GPS measurement error (C). We argue that C can be interpreted as a quality measure for movement data recorded with a GPS. If there is a strong autocorrelation between any two consecutive position estimates, they have very similar error. This error cancels out when average speed, distance or direction is calculated along the trajectory. Based on our theoretical findings we introduce a novel approach to determine C in real-world GPS movement data sampled at high frequencies. We apply our approach to pedestrian trajectories and car trajectories. We found that the measurement error in the data was strongly spatially and temporally autocorrelated and give a quality estimate of the data. Most importantly, our findings are not limited to GPS alone. The systematic bias and its implications are bound to occur in any movement data collected with absolute positioning if interpolation error can be neglected.

Digging into the history of VGI data-sets: results from a worldwide study on OpenStreetMap mapping activity

Volunteered geographic information (VGI) data-sets are characterised by heterogeneity due to influences from technical, social, environmental or economic factors. As a result, mapping progress does neither follow a spatially nor a temporally equal distribution, and thus can be hardly measured or predicted. Positively stated, heterogeneity leads to interesting VGI data-sets revealing regional peculiarities such as diverse community activities. This work proposes an approach for identifying regionally and temporally different developments with respect to mapping progress. Regional mapping progress is measured with a modified version of a previously proposed model for classifying activity stages, which has been used as foundation for a massive spatial and temporal analysis of the worldwide OpenStreetMap contributions between the years 2006 and 2013. It also allows the evaluation of rural and unpopulated areas. Results reveal that regional mapping progress heavily depends on a number of distinct influences such as geographical or legal borders, data imports, unexpected events or diverse community developments. The work highlights regions with distinct results by revealing individual mapping stories.

What is an Appropriate Temporal Sampling Rate to Record Floating Car Data with a GPS

Floating car data (FCD) recorded with the Global Positioning System (GPS) are an important data source for traffic research. However, FCD are subject to error, which can relate either to the accuracy of the recordings (measurement error) or to the temporal rate at which the data are sampled (interpolation error). Both errors affect movement parameters derived from the FCD, such as speed or direction, and consequently influence conclusions drawn about the movement. In this paper we combined recent findings about the autocorrelation of GPS measurement error and well-established findings from random walk theory to analyse a set of real-world FCD. First, we showed that the measurement error in the FCD was affected by positive autocorrelation. We explained why this is a quality measure of the data. Second, we evaluated four metrics to assess the influence of interpolation error. We found that interpolation error strongly affects the correct interpretation of the car’s dynamics (speed, direction), whereas its impact on the path (travelled distance, spatial location) was moderate. Based on these results we gave recommendations for recording of FCD using the GPS. Our recommendations only concern time-based sampling, change-based, location-based or event-based sampling are not discussed. The sampling approach minimizes the effects of error on movement parameters while avoiding the collection of redundant information. This is crucial for obtaining reliable results from FCD.

Estimating Completeness of VGI Datasets by Analyzing Community Activity Over Time Periods

Due to the dynamic nature and heterogeneity of Volunteered Geographic Information (VGI) datasets a crucial question isu concerned with geographic data quality. Among others, one of the main quality categories addresses data completeness. Most of the previous work tackles this question by comparing VGI datasets to external reference datasets. Although such comparisons give valuable insights, questions about the quality of the external dataset and syntactic as well as semantic differences arise. This work proposes a novel approach for internal estimation of regional data completeness of VGI datasets by analyzing the changes in community activity over time periods. It builds on empirical evidence that completeness of selected feature classes in distinct geographical regions may only be achieved when community activity in the selected region runs through a well-defined sequence of activity stages beginning at the start stage, continuing with some years of growth and finally reaching saturation. For the retrospective calculation of activity stages, the annual shares of new features in combination with empirically founded heuristic rules for stage transitions are used. As a proof-of-concept the approach is applied to the OpenStreetMap History dataset by analyzing activity stages for 12 representative metropolitan areas. Results give empirical evidence that reaching the saturation stage is an adequate indication for a certain degree of data completeness in the selected regions. Results also show similarities and differences of community activity in the different cities, revealing that community activity stages follow similar rules but with significant temporal variances.

Deriving driver-centric travel information by mining delay patterns from single GPS trajectories

Crowd-sourcing approaches for generating accurate real time travel information for road networks is promising but still challenging. For example, travel speeds, even if derived from highly sampled GPS trajectories, have limitations in their interpretability for more sophisticated travel information such as traffic-related delays or level of service (LOS) information. The proposed algorithm in this work analyzes the flow characteristics of individual vehicles by deriving and classifying delays into LOS relevant (e.g. queuing traffic) and LOS non-relevant delay patterns (e.g. stopping at a crosswalk). In contrast to other approaches, the proposed algorithm works on single GPS trajectories collected from individual vehicles (e.g. floating car data - FCD), without the necessity to average travel speeds or travel times of multiple vehicles for reliable LOS estimation. Applied to sample GPS trajectories from test drives the algorithm reaches an overall recognition rate of 82.0% for delay classes slight delay, massive delay, single stop, and multi stops. The recognized delay patterns are capable to distinguish between LOS relevant and LOS non-relevant delays at high accuracy for subsequent delay and LOS information. The recognition of LOS non-relevant single stops reaches a rate close to 100.0%.

Motion pattern analysis enabling accurate travel mode detection from GPS data only

Travel modes are one of the crucial pieces of information to characterize ones travel behavior. In recent years several approaches of mode detection from GPS data have been proposed. The approach presented in this paper uses machine learning to evaluate a set of GPS-based features for their ability to recognize the common modes walk, bicycle, car, bus, and train. The proposed features describe motion characteristics from GPS-trajectories by relative frequencies. Compared to previous work the proposed feature set leads to higher average recognition rates around 92% without relying on additional GIS or real-time information. The evaluation compares detection rates from multilayer perceptrons, logistic model trees, and C4.5 decision trees and is complemented by an evolutionary feature selection for selecting the most beneficial feature subsets leading to the best quality gain. In contrast to other research, this study uses a comparatively large set of 400 GPS trajectories which have been recorded in rural and urban European areas. Results contribute to a higher reliability as well as a broader applicability of GPS-only travel mode detection.

Collecting floating car data with smartphones: results from a field trial in Austria

Abstract This work reports on results from a field trial regarding the collection of floating car data with smartphones in Austria. The field trial has been conducted within Austria’s National Floating Car Data Testbed pursuing the goal to test different aspects of floating car data technology for traffic data collection, traffic state estimation and traffic prediction. The test bed collects, processes and analyses FCD from several thousand vehicles. The field trial for smartphone-based data collection has been conducted within the Federal State of Salzburg covering 1500 kilometres of major road network. Between the launch of the Android-based smartphone application in March 2014 and the end of the field trial in February 2015, the application has been downloaded by more than 2100 users. One year after launch the app is still installed on 650 devices and attracts around 15 users daily. The work gives insights into the application’s concepts and discusses app usage statistics, usage patterns and user feedback in the context of community-driven traffic data collection. On the one hand, results from the field trial confirm that community-driven traffic data collection is still not a phenomenon of the masses due to various challenges discussed throughout the work. On the other hand, results contribute to a deeper understanding of community-driven data collection in the traffic domain and help to learn for future trials.

Evaluating GPS sampling rates for pedestrian assistant systems

Abstract In recent years, assistant systems have come to widespread use and support people in various situations e.g. in getting from A to B. For quite a time also assistant systems with special attention to older people have been developed. For example, in case of cognitive impairments where autonomous living indoors as well as outdoors is affected, assistant systems can be valuable aids. First attempts for outdoor assistance with GPS-based location systems offering the possibility to define geo-fences for raising an alert if a known area is left have been made. The quality of these systems is largely dependent on the precision of localisation which among others is influenced by the sampling rate. This paper reports on an empirical study under real world conditions to determine a suitable GPS sampling rate for movement analysis of (cognitively impaired) pedestrians. The work considers GPS measurement and interpolation errors as well as track losses as a result of low sampling rates. For the study, GPS data for different environmental settings and movement scenarios for sampling rates of 1, 2, 3, 4, 5, 10, 15, 20 and 25 s has been collected. The impact of sampling rates on movement parameters like track length and speed has been empirically measured. Additionally, the influence of smoothing approaches on data quality and whether downsampling of data has the same effect as recording with corresponding lower sampling rate has been studied. Results show that across all tested scenarios a sampling rate of 3–5 s seems to be appropriate with respect to speed and track length. Additionally, it can be argued that smoothing improves data quality of highly sampled data (up to 4 s). With downsampling, outliers are less in comparison to data sampled at the corresponding sampling rate.

A model to estimate and interpret the energy-efficiency of movement patterns in urban road traffic

Abstract Urban road traffic is highly dynamic. Traffic conditions vary in time and with location and so do the movement patterns of individual road users. In this article, a movement pattern is the behaviour of a car when traversing a road link in an urban road network. A movement pattern can be recorded with a global navigation satellite system (GNSS), such as the Global Positioning System (GPS). A movement pattern has a specific energy-efficiency, which is a measure of how fuel-intensively the car is moving. For example, a car driving uniformly at medium speed consumes little fuel and, therefore, is energy-efficient, whereas stop-and-go driving consumes much fuel and is energy-inefficient. In this article we introduce a model to estimate the energy-efficiency of movement patterns in urban road traffic from GNSS data. First, we derived statistical features about the cars movement along the road. Then, we compared these to fuel consumption data from the cars controller area network (CAN) bus, normalized to the cars overall range of fuel consumption. We identified the optimal feature set for prediction. With the optimal feature set we trained, tested and verified a model to estimate energy-efficiency, with the fuel consumption serving as ground truth. Existing fuel consumption models usually view movement as a snapshot. Thus, the behaviour of the car remains unknown that causes a movement pattern to be energy-efficient or energy-inefficient. Our model views movement as a process and allows to interpret this process. A movement pattern can, for example, be energy-inefficient because the car is driving in stop-and-go traffic, because it is travelling at high speed, or because it is accelerating. Our model allows to distinguish between these different types of behaviours. Thus, it can provide new insights into the dynamics of urban road traffic and its energy-efficiency.

Generation of Meaningful Location References for Referencing Traffic Information to Road Networks Using Qualitative Spatial Concepts

Location referencing systems (LRS) are a crucial requisite for referencing traffic information to a road network. In the past, several methods and standards for static or dynamic location referencing have been proposed. All of them support machine-interpretable location references but only some of them include human-interpretable concepts. If included, these references are based on pre-defined locations (e.g. as location catalogue) and often miss meaningful interlinking with road network models (e.g. locations being simply mapped to geographic coordinates). In a parallel research strand, ontological concepts for structuring road networks based on human conceptualizations of space have been proposed. So far, both methods have not been integrated. The current work closes this gap and proposes a generation process for meaningful location references on top of road networks based on qualitative spatial concepts. A prototypical implementation using OWL, Neo4J graph database and a standardized nationwide road network graph shows the practical applicability of the approach. Results indicate that the proposed approach is able to bridge the gap between existing road network models and human conceptualizations on multiple levels of abstraction.