In aircraft scenarios the proper interpretation of communication meanings is mandatory for security reasons. In particular some communications, occurring between the signalman and the pilot, rely on arm-and-hand visual signals, which can be prone to misunderstanding in some circumstances as it can be, for instance, because of low-visibility. This work intends to equip the signalman with wearable sensors, to collect data related to the signals and to interpret such data by means of a SVM classification. In such a way, the pilot can count on both his/her own evaluation and on the automatic interpretation of the visual signal (redundancy increase the safety), and all the communications can be stored for further querying (if necessary). Results indicate that the system performs with a classification accuracy as high as 94.11 ± 5.54 % to 97.67 ± 3.53 %, depending on the type of gesture examined.
CITATION STYLE
Saggio, G., Cavrini, F., & Pinto, C. A. (2017). Recognition of arm-and-hand visual signals by means of SVM to increase aircraft security. In Studies in Computational Intelligence (Vol. 669, pp. 444–461). Springer Verlag. https://doi.org/10.1007/978-3-319-48506-5_23
Mendeley helps you to discover research relevant for your work.