%0 Journal Article %T Feature fusion algorithm based on modular scalable integrated sensor behavior recognition %A Wei, Fuyin %A Xiang, Fei %A Chu, Bohao %A Noche, Bernd %J Logistics Journal : Proceedings %D 2021 %V 2021 %N 17 %@ 2192-9084 %F wei2021 %X Based on the behavior recognition model of Convolutional Neural Network, we developed a modular scalable integrated (MSI) sensor system together with a signal feature fusion algorithm. The integrated sensor system can obtain high-quality signals without having to be embedded in the body of the object and has good modular scalability and timeliness. The feature fusion algorithm improves the recognition accuracy as well as the robustness of the model. %L 620 %K Convolutional Neural Network %K Feature-Fusion %K Genauigkeit %K Robustheit %K accuracy %K feature fusion %K modular scalable integrated sensor %K modularer skalierbarer integrierter Sensor %K robustness %R 10.2195/lj_Proc_wei_en_202112_01 %U http://nbn-resolving.de/urn:nbn:de:0009-14-54471 %U http://dx.doi.org/10.2195/lj_Proc_wei_en_202112_01