Thesis supervisor: András Edelmayer
Location of studies (in Hungarian): Széchenyi István Egyetem Abbreviation of location of studies: SZE
Description of the research topic:
To achieve high level vehicle autonomy, sensor fusion and deep-learning technologies are to be seamlessly integrated into the overall vehicle architecture to efficiently aggregate and make sense of multiple data streams in real time, such as on-board sensor data és position information, other consolidated on-board data, such as digital maps and environmental data coming from cooperative vehicle communication. It was clearly recognized that the accuracy of traditional localization technologies is not sufficient for autonomous navigation and a novel map technology is needed for positioning. Traditional digital map technologies provide a more or less stable, static representation of the road networks. Dynamic map building technologies are required, however, to enrich this static set of data with dynamically varying environmental information and construct a new map representation. One has to view this map as an additional sensor, just like a camera or radar on a vehicle helping to extend the vehicles understanding of the road network far beyond the range of its on-board sensors. There needs to be technical means by which one can pool data into a single map system which can be shared with other vehicles on road for the collective benefit of automated transportation. An aggregation platform gathers and process the data provided by multiple vendors at scale. Consolidation technologies are responsible to keep this data living and dynamic having aggregated knowledge of the road network, where real-time events (traffic, accidents, construction sites, lane closures and other obstacles) are collected and represented and delivered to vehicles in real-time. Consolidated on-board data (maps, lookup tables etc.) supporting the driving situation represents this technology pillar of autonomous drive. This type of map contains a much richer description of the environment than any recent localization technologies can do. Consolidation is largely based on vehicle communications constructing a shared dynamic data asset in the mist (cf. mist vs cloud). Certain parts of the idea have been implemented in the C-ITS communications architecture (LDM) that was validated in various safety use-cases in the practice.
Required language skills: angol
Deadline for application: 2018-02-28
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Login
