|Autors: TOPALOVA, I. H., TZOKEV, A. P.|
Title: Optimization of a MLP network through choosing the appropriate input set
Keywords: MLP, supervised learning, outliers values, optimization
Abstract: The neural networks find today many applications in different kinds of real-time working systems. To obtain short execution times in machine vision systems or in real-time decision-making systems, becomes a question of first importance. Therefore, the requirements to the recognition stage in such systems in reference to reduce the reaction time grow up. In this research, a neural network MLP structure is proposed for recognition of 2D objects intended for implementation in a real-time working system, based on PLC modules. A new method for choosing the appropriate input set and optimizing the MLP structure with the aim to reduce the execution time is developed. The method is tested with different kinds of captured 2D objects for real-time work of a PLC S7-314. The achieved good results give approve that the proposed method could be applied in other real-time systems.
Вид: публикация в национален форум