This paper describes a system that incrementally constructs an increasingly accurate road map from GPS traces from a single vehicle. The resulting road map contains information about the road such as road gradient which can be used by functions in a heavy vehicle to drive more effectively. The system is supposed to run on an embedded system in a heavy vehicle and is therefore design to require as little working memory and processing time as possible.Pre- and post processing techniques that counters GPS noise, random movements and improve the quality of the road map are also described, for example tunnel estimation where GPS signals are missing. An aging method, designed for data from a single vehicle, that eventually removes closed and rarely used roads is proposed.A comparison between the constructed road map and a commercial one shows that the algorithms described creates a very accurate roadmap. The performance of the system is evaluated and it is concluded that it would be possible to run it on an embedded system in a heavyvehicle.
Incremental Self Learning Road map
2014-01-01
Hochschulschrift
Elektronische Ressource
Englisch
Road Recognition From Remote Sensing Imagery Using Incremental Learning
Online Contents | 2017
|The incremental road to higher-speed rail
IuD Bahn | 2013
|