Probabilistic fusion of rural road course estimations

Schule, Florian; Koch, Christian; Hartmann, Oliver; Schweiger, Roland; Dietmayer, Klaus

doi:10.1109/itsc.2013.6728474

Cited by 3 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Precise lateral localization is then accomplished by fusing the longitudinally mapped digital map with an optical lane recognition algorithm [3] in the camera image. In subsequent works [6], a Bayesian fusion system that performs the final road course estimation is introduced. In both systems, the road course model is not a clothoid but rather lists of connected 2D points sampling the right and left borders of the lane.…”

Section: Related Workmentioning

confidence: 99%

Robust Deep-Learning-Based Road-Prediction for Augmented Reality Navigation Systems at Night

Limmer

Förster

Baudach

et al. 2016

2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC)

Self Cite

View full text Add to dashboard Cite

This paper proposes an approach that predicts the road course from camera sensors leveraging deep learning techniques. Road pixels are identified by training a multi-scale convolutional neural network on a large number of full-scene-labeled nighttime road images including adverse weather conditions. A framework is presented that applies the proposed approach to longer distance road course estimation, which is the basis for an augmented reality navigation application. In this framework long range sensor data (radar) and data from a map database are fused with short range sensor data (camera) to produce a precise longitudinal and lateral localization and road course estimation. The proposed approach reliably detects roads with and without lane markings and thus increases the robustness and availability of road course estimations and augmented reality navigation. Evaluations on an extensive set of high precision ground truth data taken from a differential GPS and an inertial measurement unit show that the proposed approach reaches state-of-the-art performance without the limitation of requiring existing lane markings. This work was partially funded by the European Commission under the ECSEL Joint Undertaking in the scope of the DESERVE project. http://www.deserve-project.eu/ 1 M. Limmer, J. Forster, D. Baudach and R. Schweiger are with

show abstract

Section: Related Workmentioning

confidence: 99%

Robust Deep-Learning-Based Road-Prediction for Augmented Reality Navigation Systems at Night

Limmer

Förster

Baudach

et al. 2016

2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Some applications like longer distance road course estimation [1], [2], rating of pedestrian endangerment [3] or map-aided predictive curve light [4] only are allowed by the a priori knowledge about the road geometry in terms of digital street maps. For these systems, a high localization accuracy on the digital map is crucial to maximize the effectiveness of the data stored inside the map.…”

Section: Introduction a Motivationmentioning

confidence: 99%

Feature evaluation of factorized self-localization

Schule

Büchner²,

Schweiger

et al. 2014

17th International IEEE Conference on Intelligent Transportation Systems (ITSC)

Self Cite

View full text Add to dashboard Cite

Localization on a digital map is a crucial point for many advanced driver assistance systems that make use of digital map data. State of the art work mainly relies on inaccurate GPS measurements, special and costly sensors or especially attributed digital maps. In contrast to that, this work presents a localization algorithm with in-vehicle available sensors and a standard navigation map. A factorized state estimation that computes position and orientation with separate features allows accurate and reliable positioning. This work thereto proposes correlationbased features that match road curvatures of digital map and sensor data for position estimation. The vehicle orientation on the digital map can then be computed by either matching an egomotion trajectory to the digital map road geometry or by using an optical lane recognition system. Sensitive parameters of this approach are thoroughly evaluated with a highly precise DGPS reference system.

show abstract