Nowadays, personal mobility and goods transportation are undergoing a significant transformation thanks to autonomous driving. However, collisions during autonomous driving may lead to disruptive damages, even the loss of human lives. Therefore, once some collisions occur during Autonomous Driving Systems (ADS) testing, it is meaningful to explore why these collisions can occur and how to solve them. In our work, we propose an analysis approach that is able to assess the relationship between the ADS parameters and the potential collisions observed in some particular traffic situations. Specifically, the approach uses fuzzification to partition ADS parameters into different categories and a spectrum-based analysis to identify which parameter categories can cause these collisions. As ADS behaviours are complex, the collisions may be due to a single parameter or combinations of two or more parameters. For scalability, the approach performs an incremental analysis, in which first single parameters are considered, and then parameters combinations of higher-order. The approach has been applied to an industrial path planner.
ROIS-OS人文学オープンデータ共同利用センター (Center for Open Data in the Humanities / CODH)は、情報学・統計学の最新技術を用いて人文学資料(史料)を分析する「データ駆動型人文学」や、人文学研究の成果に基づき構築したデータセットを超学際的に活用する 「人文学ビッグデータ」など、オープンサイエンス時代の新しい人文学研究を展開します。