Keywords Recognition in Aircraft Cockpit based on Genetic Algorithm-Dynamic Time Warping

摘要

This paper proposed a Genetic Algorithm-based Dynamic Time Warping (GA-DTW) algorithm to recognize the keywords in aircraft cockpit communication. The proposed algorithm can accurately extract keywords related to the various flight events from a large number of flight speech records. It is greatly reduced the workload of researchers who study aviation human factors and provides valuable materials for the subsequent identification and decomposition of flight events. Experimental results show that the proposed algorithm is not only able to improve the recall rate of keywords, but also can enhance the detection accuracy of keywords.