Title :
Emergency escape methodology powered by Businesses process modeling
Author_Institution :
Faculty of Electrical Engineering, Czech Technical, University in Prague, Prague, Czech Republic
Abstract :
The navigation within the building is possible to display using the oriented graph. To find shortest way in the graph is possible to solve in the different mode. Our question is Is possible to predict the shortest way by artificial Intelligence, primarily by Perceptron neural network with one secret layer. During the test project was used algorithm Dijkstra [1] and was developed tool for automatic way generation. Its name space is cz.cvut.fel.BHHETool [2] and is possible freely disseminate in according to GNU license. The tool allows generating the shortest way by input vector for neural network. The neural network is modeled by the tool Neuroph [3]. This article deals with the process and practical verification the possible prediction and to measure the deviation from the algorithm tools.
Keywords :
Adaptation models; Biological system modeling; Buildings; Legged locomotion; Mathematical model; Navigation; Predictive models; Artificial Intelligence; Business Process Modeling; Capacity edge; Capacity node; Dijkstra; Emergency; Escape plan; Human being; Methodology; Neuroph; Oriented chart; Prediction; Safety;
Conference_Titel :
Intelligent Green Building and Smart Grid (IGBSG), 2014 International Conference on
Conference_Location :
Taipei, Taiwan
DOI :
10.1109/IGBSG.2014.6835278