ANS Based Submarine Simulation

摘要

This report discusses the project background, including the requirements of sucha simulation model, and the advantages of an ANS approach. The uniqueness of AWI's Optimal-Entropy Neural Network algorithm is discussed as well as its significance to the success of this project. AWI has developed an ANS to model submarine performance based on the setting of the input parameters to result in a particular performance for the submarine where the ANS specifies the Position and orientation of the submarine sometime in the future. AWI has also developed an algorithm to run the ANS in the inverse mode, namely the algorithm allows the user to specify the desired position and orientation of the submarine at some time in the future, where the ANS will then specify the controlling input parameters to reach that specified objective. The latter specifications have to be within the realm of possible requirements or else the ANS will specify a possible solution close to what was desired. The developed ANS is capable of operating in a PC environment. The results are obtained almost instantaneously on the PC. Artificial neural system, Neural network, Submarine simulation modeling.