Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

World modeling for cooperative intelligent vehicles

Publication files not online:

Author: Papp, Z. · Brown, C. · Bartels, C.
Type:article
Date:2008
Place: Piscataway, NJ, USA
Institution: TNO Industrie en Techniek
Source:2008 IEEE Intelligent Vehicles Symposium IV, 4-6 June 2008, Eindhoven, The Netherlands, 1050-1055
Identifier: 241304
ISBN: 9781424425693
Article number: No.: 4621272
Keywords: Dynamical systems · Intelligent vehicle highway systems · Software prototyping · Abstraction layers · Application developments · Complex dynamical systems · Data interpretations · Digital maps · Formal representations · High resolutions · Key elements · Onboard systems · Prototype implementations · Sensory inputs · System architectures · System decompositions · Vehicle applications

Abstract

Cooperative intelligent vehicle systems constitute a promising way to improving traffic throughput, safety and comfort. The state-of-the-art intelligent-vehicle applications usually can be described as a collection of interacting, highly autonomous, complex dynamical systems (the individual vehicles). The vehicles implement onboard a wide spectrum of (possibly cooperative) functionalities (applications). In order for an application "to do the right thing" it should have a formal representation and understanding of the relevant surrounding world to implement the proper (e.g. efficient, safe) behavior. In the proposed approach, developed for the SAFESPOT Integrated Project, the Local Dynamic Map (LDM) forms a key element of the onboard system responsible for representing and maintaining a real-time world model. From the applications' point of view, the LDM is the model of the world as known by the vehicle: it contains objects characterized by attributes, uncertainties and relevant relationships between objects. The world model in the LDM is created based on high resolution digital maps, sensory inputs and communication. The LDM implements an abstraction layer between the data interpretation and the higher level behavioral functions. The paper describes the motivation behind the LDM based system decomposition approach, introduces the basic concepts of the world modeling dedicated to cooperative vehicle systems and summarizes the main features of a prototype implementation. The consequences of the LDM based system architecture on application development and testing are also presented.