@inproceedings {pub3315,
	title = {Driving Situation Analysis with Relational Local Dynamic Maps (R-LDM)
},
	author = {Julian Eggert AND Daniela Aguirre Salazar AND Tim Puphal AND Benedict Flade},
	year = {2017},
	month = {September},
	abstract = {In the automotive domain, the concept of Local Dynamic Maps (LDM) is currently used to describe a central storage place where static, quasi-static and dynamic road information is integrated by means of a common coordinate reference. In this paper, we argue that future ADAS and AD applications concentrating on driving situation analysis will require, instead of a layered, a fully interconnected graph-based approach, which we propose here in form of a Relational Local Dynamic Map (R-LDM). The graph-based approach lends itself for the creation of consistent world models with richer interconnected structures on spatial, temporal and semantic levels. We explain the design and explore the capabilities of the R-LDM in one application for camera-to-map alignment for lane-detailed localization, and in another application for the preparation of predicted trajectories for risk-based behavior evaluation.
},
	publisher = {Society of Automotive Engineers of Japan},
	booktitle = {FAST-zero Symposium 2017}
}