A real-time autonomous highway accident detection model based on big data processing and computational intelligence Article Swipe
YOU?
·
· 2016
· Open Access
·
· DOI: https://doi.org/10.1109/bigdata.2016.7840798
· OA: W2583901265
Due to increasing urban population and growing number of motor vehicles,\ntraffic congestion is becoming a major problem of the 21st century. One of the\nmain reasons behind traffic congestion is accidents which can not only result\nin casualties and losses for the participants, but also in wasted and lost time\nfor the others that are stuck behind the wheels. Early detection of an accident\ncan save lives, provides quicker road openings, hence decreases wasted time and\nresources, and increases efficiency. In this study, we propose a preliminary\nreal-time autonomous accident-detection system based on computational\nintelligence techniques. Istanbul City traffic-flow data for the year 2015 from\nvarious sensor locations are populated using big data processing methodologies.\nThe extracted features are then fed into a nearest neighbor model, a regression\ntree, and a feed-forward neural network model. For the output, the possibility\nof an occurrence of an accident is predicted. The results indicate that even\nthough the number of false alarms dominates the real accident cases, the system\ncan still provide useful information that can be used for status verification\nand early reaction to possible accidents.\n
