Car park mapping with simultaneous localisation and mapping (SLAM)

摘要: One of the challenges in automation vehicles takes place car park. Automatic must be able to localize itself accurately and build a map park real time. This capability will extend range possible applications including tracking lots availability, automatic navigation parking name few. In this project, we took on task localizing an vehicle building project within INRIA Rhone-Alpes CyCab with Sick laser scanner. A key feature is that it works only scanners retrieve position orientations With detected as landmarks, performs localization at same problem commonly known literature Simultaneous Localization Mapping(SLAM). The detection based work previous DEA student (Lorieux) extensions were proposed increase reliability detection. SLAM algorithm chosen FastSLAM algorithm. adapted context. However, gives set hypotheses. Hence, construction method merge different hypotheses together form one single final map. Experiments data conducted perform tasks required. experiments also revealed weaknesses system approaches directions for further research suggested.