The accurate classification of moving object in urban traffic scenarios is a key element for safe decision-making of intelligent vehicles. Multi-sensor approaches are typically based on features or specific objects models, which results in either an intrinsic lack of robustness or a significant design complexity. This paper proposes to take advantage of a Bayesian occupancy framework for perception, introducing a classifier combining grid-based footprints and speed estimations. The preliminary results obtained with a Lidar under a very realistic simulation framework are very promising.
Footprint-based classification of road moving objects using occupancy grids
2017 IEEE Intelligent Vehicles Symposium (IV) ; 1052-1057
01.06.2017
1693254 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Footprint-Based Classification of Road Moving Objects Using Occupancy Grids
British Library Conference Proceedings | 2017
|Road course estimation in occupancy grids
IEEE | 2010
|Road Course Estimation in Occupancy Grids, pp. 412-417
British Library Conference Proceedings | 2010
|Learning Occupancy Grids of Non-Stationary Objects with Mobile Robots
Springer Verlag | 2003
|