Shadowed type 2 fuzzy-based Markov model to predict shortest path with optimized waiting time

Abstract

Recently, the traffic network exhibits a very critical situation due to the speedy rise of urbanization and population growth. This paper suggests a better solution for such traffic issues via delay-optimized Shortest Path Prediction (SPP) method. Even though Shadowed Type 2 (ST2) fuzzy logic works well for delay optimization with uncertain data, it causes a rise in fuzzy partitioning complexity. This motivates to development Shadowed Type 2 Fuzzy Markov (ST2FM) scheme for accurate prediction of the shortest path. In ST2FM, waiting for time optimization performed at rush junction based on ST2 fuzzy rules. Optimized path detail is periodically updated in the Transition Probability Matrix of the Markov model for SPP. Thus, ST2FM helps a node to easily identify the shortest path to reach the destination without waiting at traffic junctions. The absence of fuzzy partitioning and the use of Markov prediction greatly reduce the computational complexity of ST2FM. Matlab 2016a working environment is utilized for research implementation and results are compared with ST2 fuzzy, Interval Type 2 fuzzy and Fuzzy-based Convolution Neural Network. From this analysis, the proposed work demonstrates 96% of prediction accuracy with less error than existing works.