4-4-6 
6 DATA PROCESSING 
Due to the hilly areas^ satellite masking effects provide 
disturbances up to 70m in the GPS-signals. Between masking 
effects the shape of the track is measured precisely. Useful GPS- 
measurement can be selected by plausibility checks. The 
correction of heading changes and therefore the shape of the 
measured track can be compared to differential odometer and 
odometer and GPS-measurements. Nevertheless the problem of 
the datum cannot be chosen by this way. 
Curvature in 1/km 
This picture results in 2 basic conclusions: 
- for short measurement tracks with less than 3 curvature zeros, 
the map matching strategy is unreliable and can fail. 
- With growing distance driven on a certain track the reliability 
of positioning is nearly independent of the available sensor. 
Odometer and GPS only odometer or DGPS differ only by 
approx. m in accuracy of positioning relatively to the 
length of the road network. This is an important result for 
future car navigation systems. It means odometer and simple 
GPS-measurement techniques can be used with nearly the 
same accuracy of DGPS-sensors for car navigation purposes. 
These theoretical and practical results will be used in future 
car navigation systems in Germany. 
8 SUMMARY 
In different examples for curvature pattern analysis efficient car 
navigation computation strategies based on map matching are 
developed. In case of board autonomous car navigation systems 
the consequent solution based on multi-sensor approaches can 
lead to extremely cost reductions compared to functional 
availability and operability of these systems. Nevertheless only 
GPS-based solutions lead to system failures by masking and 
second-way-affects - the physical system limits of GPS in urban 
areas. 
Picture 9 chose a curvature matching process for the related road 
example. An accuracy of 50-70cm can be provided in this case. 
For practical approaches it is of tremendous importance that due 
to continuous map matching compared to the digital road map in 
the curvature level drift effects of the sensors are mainly without 
any affect to the positioning result. Therefore curvature pattern, 
positioning strategies in map matching techniques lead to a new 
level of accuracy in map matching technologies especially for 
urban areas. 
7 RELIABILITY OF TRACK IDENTIFICATION 
Picture 10 shows the accuracy of different sensors combinations 
dependent on the length of the measured tracks. 
Accuracy in m 
REFERENCES 
[Schmeck, 1992] Schmeck, K.H., 1992. Computer Integrated 
Transport Logistics for European Traffic in: Proc. Intern. 
Congress on Transportation Electronics, Michigan 1992, ISBN 1- 
66091-267-7 
[Both, Davidsson, Möller, 1991] Both, M., Davidsson, F., Möller, 
K. 1991. Pan-European Freight and Fleet Management Brussels, 
Febr. 4-6, 1991, Vol. II ISBN 0-444-89043-2 
[Möhlenbrink, Mezger, 1996] Möhlenbrink, W., Mezger, K., 
1996. GPS-Anwendungen im Straßenverkehr, Spektrum der 
Wissenschaft, Januar 1996 
[Möhlenbrink, 1994] Möhlenbrink W., 1004. Integrated Traffic 
Management, 13 th Wolrd Compter Congress 94, IFIP 1994, 
Elsevier Science B.V. (North Holland) 
[Möhlenbrink, 1996] Möhlenbrink, W., 1996. 
Managementsysteme für den intermodalen Verkehr. VDE- 
Kongreß, Oktober 1996, Brauschweig 
[Möhlenbrink, 1998] Möhlenbrink, W., 1998. Traffic Guidance 
and Information Systems - New Technologies for the 
Geoinformation Market. Symposium on Geodesy for 
geotechnical and Structural Engineering, April 1998, Eisenstadt 
bei Wien 
[Zhao 1997] Zhao, Y., 1997. Vehicle Location and Navigation 
Systems, Intelligent Transportation Systems, Artech House, Inc. 
Boston, London 1997 
Picture 10: Accuracy of different sensor combinations 
[Westbrook, Turner, 1994] Westbrook, M.H., Tuner, H.D., 1994. 
Automotive Sensors, Institute of Physics Publishing, Bristol and 
Philadelphia, 1994.