Simultane Zustands- und Existenzschätzung mit chronologisch ungeordneten Sensordaten für die Fahrzeugumfelderfassung
FakultätenFakultät für Ingenieurwissenschaften und Informatik
Automated driving systems must reliably and quickly perceive the environment in which they operate. Multi-sensor fusion is a suitable means to combine the advantages of different measurement techniques. However, fusion of different sensors may lead to out-of-sequence measurements (OOSM), i.e., asynchronous measurements where the original order of the measurements is lost. High-performance out-of-sequence algorithms are therefore needed that do not depend on the order of the measurements. In addition, existence probabilities can increase the reliability of the fusion system, especially in safety critical applications. A well-known approach to include existence probabilities in tracking is the Joint Integrated Probabilistic Data Association (JIPDA) filter. This algorithm has proven very reliable in practice. However, the handling of out-of-sequence measurements in JIPDA has been an unsolved problem up until now. This work therefore derives two new algorithms that can handle out-of-sequence measurements in state estimation as well as in existence estimation with JIPDA. A prototype Mercedes test vehicle is used for sensor fusion of a monocular camera and different radar sensors. Special emphasis is put on the temporal synchronization and determination of time characteristics of all sensors. Several different methods for precise timestamping and temporal calibration are therefore developed and validated. The derived OOSM algorithms for state and existence estimation are evaluated in different test drives. The two new algorithms are shown to perform better than state-of-the-art filters. For the first time, the correct handling of out-of-sequence measurements is now possible not only in state estimation, but also in existence estimation.
Erstellung / Fertigstellung
Normierte SchlagwörterFahrerassistenzsystem [GND]
Driver assistance systems [LCSH]
Multisensor data fusion [LCSH]