Industrial Wireless Network Access Selection Based on Multilevel Fuzzy Neural Network

In order to solve the issues in the heterogeneous industrial wireless network access selection such as delay and reliability, this paper present an industrial wireless network access selection based on multilevel fuzzy neural network. The algorithm divides the input attributes into three categories: network attributes, terminal attributes, and environment attributes. It synchronously inputs them into the fuzzy neural network to obtain the candidate network scores, which effectively reduces the number of fuzzy rules and reduces the algorithm complexity. Furthermore, this paper also consider the influence of the industrial environment on network selection and quantitatively reduce the number of switching and blocking rate. The simulation experimental results show that this algorithm is able to achieve fast network access selection in noisy and occluded industrial heterogeneous wireless networks, and the blocking rate can remain low in the presence of dense terminals.