Details of Research Outputs

Road conditions are one of the significant threats to driving safety for autonomous vehicles. The perception of road conditions is challenging because they are time-sensitive and location-sensitive. The existing solution is to integrate many advanced sensors into an autonomous vehicle to sense the surrounding in a real-Time manner. The consequence is that data collection and integration could also be problematic because they are from various sources; some may not be accurate or consistent. This study proposes a P2P-based algorithm to dynamically and accurately measure road conditions through Mobile Vehicular Networks without human intervention. The measured road conditions may change with time. Then we propose a data management algorithm to maintain their data quality at a high level. To allow vehicles to assess the risk of the road conditions quantitatively, we develop an integrated risk indicator for each identified road condition. Autonomous vehicles use the risk indicators to avoid potential troubles during path selection. The ultimate objective is to minimize the risk for a trip under the travel time constraint. We used simulations to evaluate the effectiveness of the proposed algorithms. The results showed that the proposed algorithm could achieve reasonably good measurement reliability when 85% of vehicles or above work correctly. We also simulated the efficiency of path selection. We found that an optimal path can be found in a directed graph with 100 vertices in 10ms by using an ordinary PC.