Universität Stuttgart
Permanent URI for this communityhttps://elib.uni-stuttgart.de/handle/11682/1
Browse
1 results
Search Results
Item Open Access Recovery of the geometric road design elements using low-cost sensors : a feasibility study(2009) Ogonda, Godfrey Onyango; Kleusberg, Alfred (Prof. Dr.-Ing.)Roadway inventory databases are continuously being applied in two main areas, first in the design, construction and maintenance of new and existing roads and secondly in the specific areas of intelligent transport systems such as driver assistance and mobile information systems. The initial development, storage and management of the database have therefore become the responsibility of the various road agencies around the world, which ascertain that this is done in accordance with the respective road policies and design standards in the respective states or regions. The roadway inventory database has been estimated to comprise of about 49 possible collectible categories of roadway inventory data elements such as the horizontal curve elements, the vertical alignment elements, drainage, the number of lanes and the traffic signs among others. As a result of the rapidly changing roadway environment, the challenge therefore is the up-dating of the roadway inventory database using relatively accurate and cost-effective techniques, as opposed to the mobile mapping systems usually employed in their initial development. This calls for investigation into other low-cost data collection and processing methods and devices. The aim of this study is therefore to develop a methodology for extracting the geometric road design elements from position time series derived from a set of low-cost sensors, and to assess the accuracy potential of such sensors. The report begins with a review of the basics of road design and the formulations used in the estimation of the geometric road design elements as well as a brief mention of the road design standards and guidelines. The various sensors and systems for land based positioning and navigation, with emphasis on their costs and sources of error, are discussed. An overview of the theory of curve fitting functions and criteria with specific interest in splines and their properties is also presented. The investigations, tests and data analyses carried out are based on computer simulated data as well as on field measured data collected in the vicinity of the city of Stuttgart. A set of spline-based algorithms is developed and tested for the extraction of geometric road design elements for the purpose of updating the roadway inventory database. Furthermore, a set of low-cost navigation sensors, comprising of wheel speed sensors, a single-axis micro electro-mechanical systems accelerometer and a single-axis micro electro-mechanical systems gyroscope, was built up and utilised for the acquisition of vehicle track positions. The methodologies and formulations used in the synchronisation of the measurements and the calibration of the sensors, the integration of the sensors' data based on a Kalman filter algorithm, the estimation of curvature and recovery of the geometric road design elements are presented. In summary, the tests and data analyses illustrate that - the developed spline-based algorithms successfully extract the geometric road design elements based on the assumption that the road geometry consist of straight lines, circular and transitional curves, as demonstrated by the analysis of simulated data. - the field calibration procedure of the wheel speed sensors employing a Kalman filtering process achieve an accuracy of +0.04 m in distance increments. The derived heading differences have a final cumulative deviation of +3° or a drift of about 0.0025 °/s. - the proposed set of low-cost sensors can be a viable solution as a short duration georeferencing device for mobile mapping systems in the recovery of the geometric shape of the road track, as long as (i) the vehicle is driven at low speed (about 36km/h or 10m/s) but certainly above the threshold speed of the wheel sensors, (ii) the random errors in the measurements are characterized and reduced and (iii) the pitch angle and gravitation effect are compensated for in the measured data. - the random and the systematic errors in the measured vehicle track position time series are a serious predicament to curvature estimation, which however can be reduced by the use of smoothing splines and by calibration of the measuring devices, respectively. Based on the above tests and analyses, the low-cost sensors together with the developed algorithms generated sufficiently accurate geometric road design elements for updating roadway inventory databases especially those intended for intelligent transport system applications whose accuracies are specified in metres to sub-meter. However, the recovered geometric road design elements may not be sufficiently accurate for roadway inventory databases used for road design and construction because here the accuracy requirements are in centimetre level.